Penetrating the Network of Fusarium Oxysporum Populations, Banana Cultivars and Graminoids

Promotor: Prof. Dr. Ir. Monica Höfte Laboratory of Phytopathology Department of Crop Protection Faculty of Bioscience Engineering Ghent University

Co-promotor: Dr. Ir. Soraya de Carvalho França Laboratory of Phytopathology Department of Crop Protection Faculty of Bioscience Engineering Ghent University Currently: Biobest NV.

Dean: Prof. Dr. Ir. Marc Van Meirvenne

Rector: Prof. Dr. Rik Van de Walle

Penetrating the network of Fusarium oxysporum populations, banana cultivars and graminoids:

Towards holistic management of Fusarium wilt in banana

Pauline DELTOUR

Thesis submitted in fulfillment of the requirements for the degree of Doctor (PhD) in Applied Biological Sciences: Agricultural Sciences Dutch translation of the title: Opklaren van interacties tussen Fusarium oxysporum populaties, banaan cultivars en grasachtigen: Op weg naar holistisch beheer van Fusarium verwelkingsziekte in banaan.

Cover illustration: Growing bananas in an agroforestry system. Cooperafloresta, Barra do Turvo, São Paulo, Brazil. Back cover: Micro- and macroconidia of Fusarium oxysporum

Cite as: Deltour P (2017). Penetrating the network of Fusarium oxysporum populations, banana cultivars and graminoids: Towards holistic management of Fusarium wilt in banana. PhD Thesis, Ghent University, Belgium.

ISBN number: 9789463570480

The author and the promotors give the authorization to consult and to copy parts of this work for personal use only. Every other use is subject to the copyright laws. Permission to reproduce any material contained in this work should be obtained from the author.

Members of the jury

Prof. Dr. Ir. Monica Höfte (Promotor) Department of Crop Protection Faculty of Bioscience Engineering, Ghent University

Dr. Ir. Soraya de Carvalho França (Co-promotor) Department of Crop Protection Faculty of Bioscience Engineering, Ghent University Currently: Biobest NV

Prof. Dr. Godelieve Gheysen (Chairwoman) Department of Molecular Biotechnology Faculty of Bioscience Engineering, Ghent University

Prof. Dr. Ir. Kris Audenaert (Secretary) Department of Applied Biosciences Faculty of Bioscience Engineering, Ghent University

Prof. Dr. Ir. Geert Haesaert Department of Applied Biosciences Faculty of Bioscience Engineering, Ghent University

Prof. Dr. Ir. Bart Lievens Department of Microbial and Molecular Systems Faculty of Engineering Technology, KU Leuven

Dr. Ir. Inge Van den Bergh Scientist and ProMusa Coordinator Bioversity International, Leuven, Belgium

Prof Dr. Christian Steinberg INRA – UMR Agroécologie Dijon, France

Dankwoord

Planten groeien op het veld en PhD’s groeien zo in het labo. De boer plant, maar geen plant groeit zonder input van water, zonlicht, nutriënten. Zo ook schrijft de doctoraatsstudent de thesis, maar is dit werk vrucht van de input van talloze andere personen, die ik graag wil bedanken. Monica, bedankt om mijn doctoraat te laten groeien in jouw lab. Het was vruchtbare grond, want dit boekje is uiteindelijk geoogst en vormt in mijn ogen een mooie vrucht. Soraya, het zaadje van dit doctoraat is gekiemd in jouw handen en van begin tot eind heb je de groei met zorg opgevolgd. Ik bewonder jouw onderzoeksstiel en wil je graag bedanken voor jouw onmisbare bijdrage.

Beste labogenootjes, jullie waren de zon en de regen voor mijn doctoraat. Onontbeerlijk. Mijn doctoraat is gegroeid op jullie kennis, vaardigheid en lach. Want ja, planten groeien beter als je er tegen praat en doctoraten ook. Allemaal hebben jullie bijgedragen aan het eindresultaat: dankuwel! Speciaal dankuwel aan Ilse (onze Mater Familias van het labo), Nadia (onze Kafka-bestrijdster), Huang (de PCR-mogelijkmaker), Silke (Partner-in-crime), Nathalie (Pure Happiness), Lien B (Fylogenetische boom-chirurge), Feyi (waardevolle tweedemeninggeefster) en Lien T (mijn linkerhand). Ellen, Filip en Lisa, als thesisstudenten hebben jullie mijn doctoraat extra bemest, waarvoor dank. Ook speciaal danku aan de iedereen die met het Phytopathology-Running-Team topprestaties heeft neergezet (Saman, Zhong, Vincent, Jolien, Nathalie, Evelien, Jasper, Andrea, Jolien, Lana, Evalyne, Silke, Osvaldo, Eva and Olumide). Dat labo is een sportieve bende, nietwaar?

Mijn doctoraat is een aantal keren verpot geweest. In de Braziliaanse bodem kende mijn PhD de belangrijkste groeischeut en daarom wil ik graag iedereen bedanken die ginder aan dit werk heeft bijgedragen. Olinto, muito obrigada por abrir as portas do seu laboratório e por toda ajuda. O seu laboratório e o povo ligado com esse laboratório (Ponês, Deco, Priscilla, Alexandre…) criaram a casa propícia para este trabalho. Irene e Dadinho, sem vocês esta tesa não existiria. Dadinho e Cida, muito obrigada pela hospidalidade no ‘sítio alternativo’. Suas observações foram a base deste trabalho. Irene, também muito obrigada pela hospidalidade e energia. Muito mais pessoas contibuiram (seja direitamente ou indireitamente). Obrigada a: as meninas da casa (Lara, Ana e Karina), Ivo e Alice, Alessandro Fortunato, Renata, Prof Salomão, os funcionários das estufas, Amanda, o CTA, os agricultores Custódio, Sebastião, Geraldo, Zim e Mateus.

iii

Een zijscheut van mijn doctoraat is verpot geweest naar Leuven. Bedankt aan Wim van den Ende, Łukasz Paweł Tarkowski, Rudy Vergauwen en Tom Struyf voor de hulp bij de suikeranalyse. Een andere zijscheut vond vruchtbare grond aan het ILVO waar Jane Debode en Pieter Cremelie hun deskundige vaardigheid lieten botvieren op de scheut. Ook binnen de faculteit kende mijn PhD verschillende groeikamers. Bedankt aan Stefaan de Neve, Marie- Christine Van Labeke, Christophe Petit, Mieke, Frieke van Collie en Katia Van Nieuland.

Uiteindelijk waren de vruchten rijp en kon er geoogst worden. Hoera! De post-harvest treatment van de juryleden zorgde voor de verbeterde kwaliteit van de vrucht. Bedankt aan Inge Van den Bergh, Bart Lievens, Christian Steinberg, Geert Haesaert, Kris Audenaert en voorzitster Godelieve Geysen. Het FWO wil ik bedanken voor de financiering van al het plantgoed, de grond, de input en de werkuren.

Ons ma en ons pa wil ik graag bedanken voor de pacht van bodem uit Denderbelle voor de experimenten. De put is toch niet te diep? Vrienden, huisgenoten, jullie waren de mycorrhiza voor mijn doctoraat: zeer bevorderlijk. Planten varen wel bij de aanwezigheid van PGPB: Plant Growth Promoting Bacteria. Ook mijn doctoraat floreerde op basis van een PGPB, namelijk de PhD Growth Promoting Bruno. Muito obrigada, seu Zito, meu PGPB.

Contents

Members of the jury ...... i Dankwoord ...... iii Contents ...... v List of Abbreviations ...... ix Chapter 1: General introduction ...... 1 1.1 Introduction ...... 2 1.2 Banana ...... 2 1.2.1 Taxonomy and origin ...... 2 1.2.2 Plant morphology ...... 3 1.2.3 Production...... 4 1.3 Fusarium wilt ...... 6 1.3.1 Fusarium oxysporum ...... 6 1.3.2 Historical impact of Fusarium wilt on banana ...... 16 1.3.3 Management of Fusarium wilt ...... 18 1.4 The soil ecosystem: a complex of interactions ...... 21 1.5 Problem statement and thesis outline ...... 24 Chapter 2: Disease suppressiveness to Fusarium wilt of banana in an agroforestry system: Influence of soil characteristics and plant community ...... 27 2.1 Introduction ...... 29 2.2 Material and methods ...... 31 2.2.1 Studied site ...... 31 2.2.2 Isolation, pathogenicity and race identification of Foc ...... 33 2.2.3 Soil sampling and assessment of disease suppression ...... 34 2.2.4 Characterization of the patches ...... 35 2.3 Results ...... 37 2.3.1 Isolation, pathogenicity and race identification of Foc ...... 37 2.3.2 Assessment of disease suppression ...... 38 2.3.3 Soil physical-chemical properties ...... 40 2.3.4 Vegetation parameters ...... 41 2.3.5 Soil biotic properties ...... 44 2.4 Discussion ...... 45 2.5 Acknowledgements ...... 48 2.6 Supplementary material ...... 48

Chapter 3: Comparative analysis of pathogenic and non-pathogenic Fusarium oxysporum populations associated with banana on a farm in Minas Gerais ...... 53 3.1 Introduction ...... 55 3.2 Material and methods ...... 57 3.2.1 F. oxysporum isolates ...... 57 3.2.2 Pathogenicity test on banana ...... 58 3.2.3 Pathogenicity test on tomato ...... 58 3.2.4 DNA extraction, PCR and sequencing ...... 59 3.2.5 Data analysis ...... 59 3.2.6 Carbon-utilization pattern ...... 60 3.3 Results ...... 61 3.3.1 Isolation of F. oxysporum and pathogenicity on banana ...... 61 3.3.2 Population diversity based on EF-1α and IGS ...... 61 3.3.3 SIX genes ...... 67 3.3.4 Carbon-utilization pattern ...... 72 3.4 Discussion ...... 73 3.5 Acknowledgements ...... 76 Chapter 4: Interaction between Fusarium oxysporum f. sp. cubense and non-pathogenic Fusarium oxysporum in graminoids ...... 77 4.1 Introduction ...... 79 4.2 Material and methods ...... 80 4.2.1 Fungal isolates and inoculum preparation ...... 80 4.2.2 Colonization of C. iria and B. decumbens by F. oxysporum isolates ...... 81 4.2.3 Primer design and qPCR ...... 83 4.2.4 Data analysis ...... 85 4.2.5 Phenotyping of F. oxysporum isolates ...... 86 4.3 Results ...... 87 4.3.1 Colonization by F. oxysporum isolates ...... 87 4.3.2 Phenotyping of the F. oxysporum isolates ...... 94 4.4 Discussion ...... 97 4.5 Conclusion ...... 101 4.6 Acknowledgements ...... 101 4.7 Supplementary material ...... 101 Chapter 5: Interaction of banana cultivars with Fusarium oxysporum f. sp cubense and non- pathogenic F. oxysporum ...... 103 5.1 Introduction ...... 105 5.2 Material and methods ...... 107 5.2.1 Fungal isolates and inoculum ...... 107 5.2.2 Root and soil colonization of F. oxysporum isolates ...... 108

5.2.3 Root exudate collection ...... 109 5.2.4 Germination of microconidia in root exudates ...... 110 5.2.5 Chemical composition of the root exudates ...... 110 5.3 Results ...... 111 5.3.1 Interaction between Foc and banana cultivars ...... 111 5.3.2 Interaction of banana cultivars with non-pathogenic F. oxysporum isolates ...... 113 5.3.3 Influence of non-pathogenic F.oxysporum on the interaction Foc-banana cultivars ...... 114 5.3.4 Root exudates ...... 115 5.4 Discussion ...... 121 5.5 Acknowledgements ...... 125 5.6 Supplementary Material ...... 125 Chapter 6: General conclusions and future perspectives ...... 129 6.1 General conclusions ...... 130 6.2 Future perspectives ...... 135 References ...... 141 Summary ...... 159 Samenvatting ...... 163 Curriculum Vitae ...... 167

vii

viii

List of Abbreviations

AU-p Approximately unbiased p-values BBTV Banana bunchy top virus bp Base pairs BSV Banana streak viruses CFU Colony forming units Col O Colatina Ouro Ct Threshold cycle CTA-ZM Centre for Alternative Technologies of Zona da Mata cv. Cultivar cvs. Cultivars D Cav Dwarf Cavendish DGGE Denaturing gradient gel electrophoresis EC Electric conductivity EF-1α Elongation factor-1α ET Ethylene ETI Effector triggered immunity FHIA Fundácion Hondureña de Investigación Agrícola Foc Fusarium oxysporum forma specialis cubense Fol Fusarium oxysporum forma specialis lycopersici FOSC Fusarium oxysporum species complex f. sp. forma specialis GCTCV Giant Cavendish Tissue Culture Variants GMO Genetically modified organisms HGT Horizontal gene transfer HST honest significant difference IGS Intergenic spacer indels Insertions and deletions ITC International Musa Germplasm Transit Centre JA Jasmonic acid LB Luria-Bertani agar LOQ Limit of quantification MAD Median absolute deviation MAMP Microbe-associated molecular patterns MG Minas Gerais MLST Multi locus sequence type MTI MAMP triggered immunity ng nanogram NGO Non-governamental organisation np-Fox non-pathogenic Fusarium oxysporum OC Organic carbon OD Optical Density OTU Operational taxonomic units

PAL Phenylalanine Ammonia-lyase PCR Polymerase Chain Reaction PDA Potato Dextrose Agar pg picogram PGPR Plant Growth Promoting Rhizobacteria PIF Produção Integrada de Frutas qPCR Quantitative Polymerase Chain Reaction QRL Quantitative Resistance Loci RAUDPC Relative area under the disease progress curve RDA Redundancy analysis RDS Rhizome discoloration score ROS Reactive Oxygen Species RQ Research Question SA Salicylic acid SIX Secreted in Xylem SNP Single nucleotide polymorphism sp. Species St Sequence type ST4 F. oxysporum f. sp. cubense subtropical race 4 TN Total Nitrogen TR4 F. oxysporum f. sp. cubense tropical race 4 VCG Vegetative compatibility group

Chapter 1

General introduction

Chapter 1

1.1 Introduction

Banana is an important food and cash crop, but its production is seriously constrained by Fusarium wilt. This disease is caused by Fusarium oxysporum f. sp. cubense, a soil-borne fungus which is very persistent in the field. No simple measures appear sufficient to control the disease and its complexity raises the need for integrated and holistic management. This thesis started from a case study of a Brazilian farmer who experienced considerable variability of the disease on his farm. This case was studied to gain inspiration on which potential measures could be part of effective Fusarium wilt management. Interesting field observations were translated in research questions which aimed to increase the understanding on the pathogen’s ecology. Knowledge on the behavior of the pathogen is key to develop efficient control measures.

1.2 Banana

1.2.1 Taxonomy and origin

Edible bananas belong to the genus Musa. This genus contains about 70 forest dwelling species in the region from India to the Pacific and Nepal to Australia. Wild plants are generally diploid. Varieties of the species Musa acuminata were domesticated and selected for sterility and parthenocarpy to decrease seed production and to produce sufficient pulp, respectively. The species M. acuminata is subdivided in subspecies due to the high intra- specific variability. Further variability in banana cultivars was obtained by crosses with different M. acuminata (A genome) subspecies or other Musa species: M. balbisiana (B genome), M. schizocarpa (S genome) and M. textilis (T genome) (Sardos et al., 2016). The S and T genome are rarely present in modern cultivars. The intra-specific variability of other Musa species is yet unidentified (Perrier et al., 2011; Sardos et al., 2016). Banana cultivars are classified in genomic groups depending on the genomes found in the hybrids and can be di- (AA, AB...), tri- (AAA, AAB, AAT…) and tetraploid (AAAB…). Genomic groups are further subdivided in subgroups based on the plant appearance and genetic origin. Cultivars belonging to the same subgroup are related to each other through a series of mutations after a single sexual event and generally display highly similar reactions to pests and diseases (Sardos et al., 2016; Vézina, 2015). Up to 1000 cultivars are grown, belonging to a limited

General Introduction number of subgroups (36 subgroups with A and B genome have been described on the MGIS database for banana accessions (MGIS, 2017)). Tripoids are generally prefered for their higher plant vigour, low seed number and fruit growth rate compared to diploids. The harvest of tetrapoids clones is often more difficult due to their larger pseudostems (MusaNet, 2016).

1.2.2 Plant morphology

Banana is a big herbaceous monocotyledon. Its stem is a short and below ground surface located rhizome or corm with orthotropic growth. The apparent stem, or pseudostem, consists of tightly enclosed leaf sheaths. At the top of the plant, the leaf sheath narrows and forms the petiole which carriers the leaf blade. After the production of on average 32 to 38 leaves, the stem apex shifts from leaf production to production of flowers and bracts. Eleven weeks after flower onset, the bunch appears on the floral stem. The female flowers, which have a bract between every two rows, appear first. These two flower rows generally develop into a hand of fruits. Lower on the floral stem, the male flowers develop, each separated by a bract. Sometimes, a third type of flowers, hermaphrodite flowers develop between the male and female flowers (Promusa, 2016; Swennen, 2004).

Roots develop in clusters of three to four from the ring of Mangin, which separates the rhizome cortex from the central cylinder of the rhizome. Primary roots branch into secondary and tertiary roots. Buds on the rhizome can develop in lateral shoots, also called suckers. A maiden sucker is a big sucker with mature leaves. A peeper is a coned shaped sucker without leaves, which develops into a sword sucker after the appearance of leaf scales with laminae. All suckers originating from the same corm are called a mat (Promusa, 2016; Swennen, 2004). The morphology of the banana plant is illustrated in Figure 1.1.

Chapter 1

Figure 1.1: The banana plant. Adapted from Simmonds (1959) and (http://www.uq.edu.au/_School_Science_Lessons/BaProj.html, accessed on 16 dec 2016).

1.2.3 Production

Banana is produced in almost all tropical and some subtropical areas of the world. In 2014, 133 million tons banana were produced in total, comprising 41% cooking bananas and plantains and 59% dessert bananas. About 90% of the total banana production was traded and consumed locally (FAO statistics; Lescot, 2015). Exported bananas belong practically exclusively to the subgroup Cavendish, but they also have an important position on local markets. Of the total dessert banana production 79% are Cavendish cultivars (Lescot, 2015).

Production systems differ in scale, going from backyard and compound gardens to plantations. Also the management practices are diverse, going from extensive, in backyards, to very intensive in plantations. Plantation bananas are frequently desuckered to maintain only the maiden sucker, replanted (depending on soil conditions), heavily fertilized and protected by agrochemicals. Planting material can be obtained from both vegetatively propagated suckers on the field and in-vitro propagated plantlets. The latter often is first

General Introduction choice as it can be certified as disease free, but is mostly too expensive for small scale farmers. Mixed farming systems in the tropics often include banana, since it is a crop that requires minimal attention and delivers food and income throughout the year (Van der Veken, 2010).

Banana is the second most important fruit in Brazil, after orange. Banana is produced in all states in Brazil, with the highest production in São Paulo, Bahia, Pará, Santa Catarina and Minas Gerais (Alves, 2009; Costa et al., 2015). Total banana production in Brazil comprised 5.16% of the total world production in 2014 and 98.8% was traded on domestic markets (Lescot, 2015). Only in India, China, The Philippines, Uganda and Ecuador, more bananas were produced in 2014 than in Brazil. The Philippines and Ecuador are the biggest exporters, respectively for Asian and North-American/European markets (FAO, 2014; Lescot, 2015).

In Brazil, banana is grown by small scale, medium scale and large scale farmers. However, the majority of the production is done by family farmers and on small scale (Moreira, 1999). Both dessert banana and cooking bananas are cultivated, accounting for 92.9% and 7.1% of the production, respectively (Lescot, 2015). The most popular cultivars in Brazil have genomic group AAB and belong to the cultivars Prata, Prata Anã, Pacovan, Maçã, Terra and Terra d’Angola. Also Cavendish cultivars (AAA) occupy an important position and various other cultivars are being cultivated, such as cv. Ouro (AA) (Borges et al., 2009). Figure 1.2 illustrates the classification options offered to Brazilian farmers for commercialization under PIF accreditation (Produção Integrada de Frutas – PIF).

Figure 1.2: Most common dessert bananas in Brazil: Cavendish cultivars (AAA, cultivars: Nanica, Nanicão, Grand Naine), cultivar Ouro (Sucrier subgroup, AA), type cultivar Maçã (Silk subgroup, AAB, culitvars: Maçã, Mysore, Thap Maeo) and type cultivar Prata (Pome subgroup, AAB, cultivars: Prata, Prata Anã, Pacovan, Branca) (PBMH & PIF, 2006)

Chapter 1

1.3 Fusarium wilt

Banana production is challenged by many biological stressors. Banana weevils and nematodes, such as Pratylenchus coffeae and Radopholus simili, are the most aggressive pests on banana. Diseases on banana can be caused by viruses, such as the Banana streak viruses (BSV) and Banana bunchy top virus (BBTV), by bacteria, such as Xanthomonas campestris pv. musacearum which causes Xanthomonas wilt and by fungi, such as Pseudocercospora fijiensis and P. musicola, the causal agent of black and yellow sigatoka, respectively. Another important fungal pathogen on banana is Fusarium oxysporum f. sp. cubense (Foc), which causes Fusarium wilt, popularly known as the Panama disease. Foc is one of the most intricate and challenging pathogens that affect banana. Below, we introduce Fusarium oxysporum f. sp. cubense, the Fusarium wilt disease in its current and historical context and management options.

1.3.1 Fusarium oxysporum

The Fusarium oxysporum species complex belongs to the fungal genus Fusarium. This genus comprises about 300 phylogenetically distinct species, most of which are soil-borne (Aoki et al., 2014). Most plant pathogens of the Fusarium genus are nested in four species complexes: F. fujikuroi (bakanae disease on rice, ear rot in maize, pitch cancer in pine, contaminants of cereals with fumonisin mycotoxins), F. graminearum (head blight of barley and wheat, contaminants of cereals with trichothecene mycotoxins), F. solani (foot and root rot of diverse hosts) and F. oxysporum (vascular wilt on several hosts) (Aoki et al.,2014). Based on molecular techniques, the taxonomy of Fusarium has been modified but is still under debate (Aoki et al., 2014; Gordon and Martyn, 1997; Laurence et al., 2014). Therefore, many fungi of the Fusarium genus are grouped in species complexes in which related, but possibly different species are gathered.

F. oxysporum is the most abundant species complex of this genus and includes non- pathogenic, plant pathogenic and human pathogenic strains (di Pietro et al., 2003). Members of the Fusarium oxysporum species complex (FOSC) are found worldwide in soils. The F. oxysporum species complex groups fungal isolates with the same morphological appearance: non-septate microconidia in false heads on short monophialides, 3-septate macroconidia formed from monophiliades on branched conidiophores in sporodochia and

General Introduction chlamydospores with smooth or rough walls, formed singly or in pairs (Figure 1.3A) (Fourie et al., 2011; Nelson, 1981). Colony morphology on plate is highly variable but F. oxysporum most typically forms a whitish-purplish mycelium (Nelson, 1981) (Figure 1.3B).

F. oxysporum is best known for its members causing vascular wilt on more than 130 hosts (Aoki et al., 2014). Generally F. oxysporum pathogens are host specific, and are divided corresponding to this host specificity in formae speciales. F. oxysporum are haploid and have no known sexual cycle (Ordonez et al., 2015). Although a sexual cycle has never been observed, a considerable amount of genetic variation has been observed in F. oxysporum (di Pietro et al., 2003). To explain this diversity, alternative mechanisms have been proposed, such as transfer of active transposable elements, mutations, heterokaryosis after a parasexual cycle and horizontal gene transfer (di Pietro et al., 2003, Gordon and Martyn, 1997; Ma et al., 2010). However, sexual recombination can not be excluded as different mating-type idiomorphs have been found in F. oxysporum (Fourie et al., 2009).

A B

b c

Figure 1.3: A: Three types of spores from F. oxysporum: microconidia (a), macroconidia (b) and chlamydospores (c). B: Colony morphology of several F. oxysporum isolates grown on potato dextrose agar (PDA).

Chapter 1

1.3.1.1 Fusarium oxysporum f. sp. cubense

All F. oxysporum isolates that can cause vascular wilt on banana are grouped under forma specialis cubense. This forma specialis is named after Cuba, the origin of the first sample from which Foc was isolated (Ordonez et al., 2015). Foc is subdivided into four physiological races based on a characteristic pattern of virulence on different banana cultivars (di Pietro et al., 2003). Race 1 is widespread over almost all banana growing regions and affects, among others, Gros Michel and Silk. Race 2 has been found on all banana growing continents, but it occurrence is more restricted in comparison to race 1 (Ploetz, 1990). Race 2 can affect cooking bananas such as Silver Bluggoe. Race 4 attacks, besides all cultivars susceptible to race 1 and 2, also bananas of the Cavendish subgroup. This race is subdivided in subtropical and tropical race 4, depending on the climatic conditions in which it causes wilt on Cavendish cultivars. Subtropical race 4 (ST4) can cause wilt in Cavendish cultivars in regions where bananas suffer a period of cold stress, such as South Africa, the Canary Islands and Taiwan, but not in tropical areas. It is still under discussion if ST4 is a genuine different race, or a race 1 able to affect disease-predisposed Cavendish cultivars. Tropical race 4 (TR4) affects Cavendish cultivars under optimal, tropical conditions. It has first been described in Malaysia and Indonesia in the 1990’s and has since spread to Australia and the rest of South-East- Asia. More recently, it has been reported in other Asian countries and Mozambique (Butler, 2013; Garcia et al., 2014; Ordonez et al., 2016). Finally, race 3 is pathogenic to Heliconia spp., a genus related to Musa sp. originating from tropical South-America (Ploetz, 2015; Waite, 1963). As this race does not affect banana, it is no longer considered part of the race structure of Foc (Fourie et al., 2011).

Within Foc, about 24 different vegetative compatibility groups (VCG) and many lineages based on different genetic markers and sequence analyses have been described (Fourie et al., 2011; Koenig et al., 1997; Ploetz, 2015). The characterization based on VCG is the most used within forma specialis cubense and is based on the potential of heterokaryon formation among nitrate-non utilizing auxotrophic (nit) mutants. Isolates form a heterokaryon when sharing common alleles at vegetative (vic) or heterokaryon (het) incompatibility loci. It is assumed that isolates sharing the same vic genes would be very similar, but a single mutation at a vic gene can cause classification in different groups. Therefore, VCG classification is a useful phenotypic trait, but provides limited genetic insight. Moreover, the relation between VCG and race is complex since within a single VCG several races can be found and within a single race, members can belong to different VCG (Fourie et al., 2011; Gordon and Martyn, 1997). For example, some isolates classified as ST4 belong to the same VCG as isolates of race 1 (Ploetz and Pegg, 2000). Sequence analysis revealed that Foc

General Introduction isolates are not monophyletic (O’Donnell et al., 1998). Some Foc isolates are more closely related to other formae speciales than to other isolates of forma specialis cubense (O’Donnell et al., 1998). The division in races is practical to describe the interaction with the plant, but fails to show the genetic diversity of Foc. In contrast to the other races, TR4 appears to have originated from a single clone since very limited genetic variation has been found among TR4 isolates (Ordonez et al., 2015).

Considering the polyphyletic structure of Foc, pathogenicity appears to have several independent origins. It is proposed that a pathogenic form can be derived from originally non- pathogens by coevolution with the host or inter-isolate exchange of chromosomes or segments, mostly called horizontal gene transfer (Fourie et al., 2011; Gordon and Martyn, 1997; Michielse and Rep, 2009). The virulence of a pathogenic isolate can change through single gene mutations, which can be mediated by transposable elements, or by the loss of chromosomes or segments (Gordon and Martyn, 1997). Although no sexual state is known, the evolvement of cryptic sexual reproduction cannot be excluded (Fourie et al., 2009; Gordon and Martyn, 1997).

1.3.1.2 Fusarium wilt of banana and the disease cycle

Foc is disseminated by the use of latently infected planting material, infected tools, the movement of infested soil on wheels and boots, by irrigation or run–off water and insects, such as banana weevils and fungus gnats (Heyman, 2015; Meldrum et al., 2013; Ploetz, 2015). In presence of a trigger, such as roots of a host plant, Foc chlamydospores germinate and grow towards and over the root surface. When the pathogen reaches the root tips or elongation zones of lateral roots or root hairs, it invades the epidermal cells directly through small openings in the cell walls (Li et al., 2011a). Foc can also enter roots via wounds and cracks in the roots. During root invasion, the pathogen produces cell wall degrading enzymes (CWDEs) such as cellulases, polygalacturonases and xylanases (Swarupa et al., 2014). The pathogen grows inter- and intracellularly until reaching the xylem. There, it produces micro- and macroconidia, which are transported upwards on the water flow. The abundant sporulation blocks the vessels. Jointly with the production of toxins, such as fusaric acid, which disturbs the membrane permeability of the plant, the water balance of the plant gets seriously disturbed and the plant wilts (Dong et al., 2012). First external symptoms in banana are the yellowing of the outer leaves, which progresses to the younger leaves. The leaves may remain erect or collapse at the petiole forming a skirt around the pseudostem. Splitting of the pseudostem, pale margins on new leaves and distortions of the leaf blade may occur.

Chapter 1

Internally, vascular discoloration can be observed from the feeder roots, to the rhizome, up to the pseudostem (Figure 1.4). If fruit is produced, it shows no symptoms (Moore et al., 1995). Finally, the plant dies off, and the heavily infected plant debris contributes to the Foc inoculum in soil. Foc can survive for decades in soil in absence of a susceptible host plant. This long term survival can be attributed to the production of thick walled chlamydospores or sclerotia (Nelson, 1981). In addition, Foc can survive as saprophyte on organic matter and as endophyte of symptomless hosts (Ploetz, 2015).

A B C

Figure 1.4: Typical symptoms of Fusarium wilt of banana. A: Vascular discoloration of the rhizome of banana cv. Maçã affected by Foc (Araponga, Brazil). B: Splitting of the pseudostem of cv. Maçã (Pedra Dourada, Brazil). C: Yellowing and collapsing of the older leaves of banana affected by Fusarium wilt (Foto: Guy Blomme, Promusa: http://www.musarama.org/en/image/fusarium-wilt-leaf-symptoms-41.html).

1.3.1.3 Virulence of F. oxysporum

To cause vascular wilt in a host plant, F. oxysporum requires an arsenal of tools. Those tools are two-fold: firstly, to effectuate the infection of the plants and secondly, to overcome plant defense responses. A lot of research is ongoing to unravel the virulence of F. oxysporum. The following paragraph gives an overview of the knowledge about the virulence of several F. oxysporum formae speciales.

General Introduction

The infection process is highly regulated and includes successful recognition of the roots through host signals and root surface attachment. F. oxysporum switches from a saprophytic to an infectious lifestyle after recognition of a susceptible host environment by G-proteins. Genes involved in pathogenesis are activated via several signal transduction pathways, mainly the protein kinase A (cAMP–PKA) and mitogen-activated protein kinase (MAPK) cascades (Guo et al., 2014; Michielse and Rep, 2009). In Foc, it has been shown that genes involved in MAPK cascades are necessary for full virulence since they encode for the production of chitin, peroxidase, beauvericin and fusaric acid (Ding et al., 2015). Mycotoxins such as fusaric acid and beauvericin are often phytotoxic, while a optimal chitin composition in the fungal cell wall, and peroxidase are necessary for the establishment of the pathogen in the non-hostile plant environment. After successful attachment, F. oxysporum can colonize the root cortex and produces an array of cell wall degrading enzymes (CWDE), such as polygalacturonases, pectate lyases, xylanases and proteases to progress towards the vascular tissue (di Pietro et al., 2003; Guo et al., 2014; Sutherland et al., 2013). Generally, those enzymes individually are not necessary for full virulence, owing to functional redundancy of the enzymes (di Pietro et al., 2003; Michielse and Rep, 2009). The production of CDWE is stimulated by the presence of the specific substrate of issue, such as cellulose, and repressed by glucose (di Pietro et al., 2003).

Upon colonization of the plant, the pathogen is exposed to various plant defense mechanisms, such as physical barriers, antifungal compounds, pathogenesis related proteins, small antimicrobial peptides and reactive oxygen species (di Pietro et al., 2003; Michielse and Rep, 2009). Plants produce gels and tyloses to block the passage of the pathogen in the xylem. In addition, soluble and cell wall bound phenolic compounds will accumulate after attack, which contribute to cell wall strengthening via lignin deposition. Phenolic compounds can have an inhibitory effect on the growth and germination of F. oxysporum (Michielse and Rep, 2009). F. oxysporum senses the presence of those compounds and can respond by increasing mycotoxin production and activity of hydrolytic enzymes, such as pectinases and cellulase (Wu et al., 2008a,b). The fungus can also degrade phenolic compounds via the β-ketoadipate pathway (Michielse et al., 2009a) or oxidate them through the action of laccases (Canero and Roncero, 2008; Qi et al., 2013).

Antimicrobial compounds can be present in the plant (phytoanticipins) or can be produced after infection (phytoalexins). In tomato plants infected with F. oxysporum f. sp. lycopersici, levels of sulphur in the xylem can accumulate to fungitoxic levels (Williams et al., 2002) and the production of the saponin α-tomatin can increase. This saponin interacts with the sterols of the fungal membranes. F. oxysporum f. sp. lycopersici copes with increased α-tomatin

Chapter 1 levels by modifying the membrane sterol content and secreting tomatinase, a tomatine- degrading enzyme (Défago et al., 1983; di Pietro et al., 2003). In the Foc genome, highly similar sequences of genes involved in detoxification have been reported (Guo et al., 2014).

Another defense reaction of the plant is the production of fungal cell wall degrading enzymes, such as chitinases, β-1,3-glucanases and β-1,4-glucosidases (Alabouvette et al., 2009; Michielse and Rep, 2009; Sutherland et al., 2013). To limit the damage by those enzymes, successful pathogens can adapt their fungal cell wall. Madrid et al. (2003) showed that the production of chitin synthase V is necessary for the construction of the fungal cell wall of Fol. In addition, successful pathogens can secrete proteolytic enzymes that target chitinases (de Sain and Rep, 2015). An adaption of the fungal cell wall can also contribute to evade the plant reaction by a reduced release of cell-wall derived defense elicitors (Michielse and Rep, 2009).

Extensive research on the tomato - F. oxysporum f.sp. lycopersici (Fol) interaction has revealed that several small cysteine-rich proteins which are excreted in the xylem (SIX: secreted in xylem) are important for full virulence (de Sain and Rep, 2015). Fourteen different SIX genes have been described in Fol (Schmidt et al., 2013). It has been demonstrated that several SIX proteins contribute to virulence in Fol, however, how they do so is yet unknown (de Sain and Rep, 2015). Homologues of SIX genes have been found in several other formae speciales and proposed as putative effector genes (Chakrabarti et al., 2011; Fraser- Smith et al., 2014; Kashiwa et al., 2013; Meldrum et al., 2012; Sasaki et al., 2015; Thatcher et al., 2013). They appear to be specific for the different formae speciales and races (van Dam et al., 2016). More so, transcription factors such as Sge1, which induces the transcription of SIX1, can be also indispensable for full virulence on tomato (Michielse et al., 2009b). A recent study showed that Foc isolates with silenced Sge1 showed reduced virulence on banana (Fernandes et al., 2016). Several other transcription factors, such as Fow2 and FTF1, are important for pathogenicity in the formae speciales melonis, and phaseoli respectively (Imazaki et al., 2007; Ramos et al., 2007). However, target genes of those transcription factor are so far unknown (Michielse and Rep, 2009; Guo et al., 2015).

A comparative genome analysis of a Fol isolate with F. graminearum and F. verticillioides revealed that the Fol genome consists out of a core genome of 11 chromosomes, for which homologues were found in the two other Fusarium species, and lineage specific regions. The lineage specific part of the genome consisted out of four entire chromosomes, and two teleomeric parts of chromosomes which are part of the core genome. The lineage specific region contained mostly additional sequences, which were unique to Fol. It does not contain

General Introduction housekeeping genes, has a peculiar high density of transposable elements and is enriched in genes that could be functionally classified as secreted effectors and virulence factors, transcription factors and proteins involved in signal transduction. Most of the SIX genes are located on one of the lineage specific chromosomes (Ma et al., 2010). Other virulence factors can be located in the core genome, suggesting the integration of core with lineage specific regions in one regulatory network (Ma, 2014; Michielse et al., 2009b). Chromosomes of both the core and lineage specific part can be horizontally transferred, although the latter with higher frequency (Ma et al., 2010; Vlaardingerbroek et al., 2016). Based on the variable copy number of the SIX8 gene among different Fol strains, considerable variability within the lineage specific regions of different Fol strains can be expected (de Sain and Rep, 2015).

Studies exploring the virulence of Foc suggest similar virulence mechanisms as described above (Guo et al., 2014; Qi et al., 2013; Sutherland et al., 2013). Guo et al. (2014) compared the pathogenesis related genes of Foc race 1 and race 4. Although both races were phylogenetically very similar, Foc race 4 had a wider biochemical repertoire and more signaling pathways were activated compared to Foc race 1.

1.3.1.4 Plant defense to F. oxysporum

A complex set of genes is involved in the resistance of plants towards pathogen attacks. In short, resistance is the outcome of a three phased cascade: the recognition of the pathogen by the plant (1) activates different signal transduction pathways (2) which result in defense responses (3) (Swarupa et al., 2014). The resistance mechanisms and genes have been intensively studied in model plants like Arabidopsis, rice and tomato. Studies about banana are limited, but show that the resistance mechanism in banana is complex and not always similar as those described in the model plants (Li et al., 2012; Swarupa et al., 2014).

The recognition of the pathogen happens in two phases. Firstly, pathogen recognition receptors detect conserved microbial molecules (MAMP: microbe-associated molecular patterns), such as chitin. The basal, broad-defense activated by MAMP recognition, is known as MTI (MAMP triggered immunity) and triggers reactive oxygen species (ROS) bursts, the activation of protein kinases and massive transcriptional reprogramming (de Sain and Rep, 2015). Secondly, when pathogens break through the MTI by secretion of effector proteins, plants can recognize those effectors by specific resistance genes (R-genes) that trigger ETI (Effector triggered immunity). ETI is generally a stronger defense response compared to MTI. A single R-gene can be responsible for effective resistance. However, often resistance to

Chapter 1 necrotrophic pathogens is the result of the sum of multiple genes with minor effects, known as quantitative resistance loci (QRL) (Zhang et al., 2013b). Those genes are located in a grey zone between MTI and ETI (Poland et al., 2009).

Major resistance genes to F. oxysporum have been described in several crops, such as tomato, melon, eggplant, pea, bean, cotton (Michielse and Rep, 2009). The existence of major R-genes indicates that resistance to some formae speciales is mediated by a gene-for- gene relationship and that different races can be defined by the R gene spectrum effective against them (Michielse and Rep, 2009). In tomato, resistance mediated by R-gene I-3, is activated upon recognition of SIX1, a small cysteine-rich protein which has been renamed to Avr3 (avirulence protein 3) (See paragraph 1.3.1.3; Rep et al., 2004). Likewise, SIX4 (Avr 1) is required for I- and I-1, and SIX3 (Avr 2) for I-2 mediated disease resistance (Houterman et al., 2008, 2009). Resistance to F. oxysporum in tomato, Arabidopsis and melon is proven to be an oligogenic trait (Li et al., 2015). Resistance to Foc in banana is described as both an effect of a single gene (Ssali et al., 2013; Vikili, 1965) and quantitative disease resistance (Dita, 2015; Li et al., 2013; Miller et al., 2008), where the outcome depends on inoculum densities. The continuous range of resistance levels found in banana cultivars supports the hypothesis that resistance is generally mediated by QRLs. However, to our knowledge, no QRL in banana has been described (Li et al., 2015).

After pathogen recognition by the plant, different signal transduction pathways are activated. Plant hormones like salicyclic acid (SA), jasmonic acid (JA) and ethylene (ET) are key players in the signal transduction. SA is upregulated as reaction against biotrophic pathogens, while the latter two mediate the reaction on necrotrophes. Also plant hormones such as auxin and abscisic acid are often involved. One of the most rapid reactions after pathogen recognition is the release of reactive oxygen in an oxidative burst, which induces other downstream signaling molecules and induces hypersensitive response mediated cell death. In addition, Ca2+ is an important intracellular secondary messenger (Swarupa et al., 2014). It has been demonstrated that resistant bananas have a higher activity of ROS scavenging enzymes (Kavino et al., 2007; Li et al., 2011b). Foc is described both as a necrotrophic or hemibiotrophic pathogen (Li et al., 2012; Swarupa et al., 2014). Following the latter, during the early root invasion, Foc acts as a biotroph, but while proliferating in the xylem, Foc is a nectrotrophic fungus (Ma, 2014; Thaler et al., 2004). Both the SA and JA/ET pathways seemed to be involved in the resistance to Foc on banana (Endah et al., 2008; Li et al., 2012; Sun et al., 2013).

General Introduction

Successful recognition and signal transduction will activate several response genes such as those coding for pathogenesis related proteins, enzymes involved in cell-wall strengthening, physical occlusion of the pathogen and antioxidants. Many defense responses have already been described in paragraph 1.3.1.3. In contrast to the lack of knowledge about pathogen recognition and signal transduction in banana, the defense responses and biochemical changes of banana upon infection by Foc have been better described (Li et al., 2013; Li et al., 2015). Upon infection, tyloses and gels are formed in the xylem to block the passage of the pathogen (Swarupa et al., 2014; Vander Molen et al., 1987). In addition, cell walls are strengthened by lignification upon an accumulation of phenolic compounds (de Ascensao and Dubery, 2003; van den Berg et al., 2007; Wang et al., 2012). Fungal cell wall degrading enzymes, such as β1,3-glucanases, are produced to inhibit the pathogens entry (Jin et al., 2007). Both in resistant and susceptible banana plants a similar reaction is found upon Foc infection, but the reaction in resistant plants is timely and effective, while not in susceptible plants (Li et al., 2013, Pegg and Langdon, 1987; Swarupa et al., 2014; Yadeta and Thomma, 2013).

1.3.1.5 Non-pathogenic F. oxysporum

Besides vascular wilt pathogens, many members of the FOSC have no known pathogenicity on plants and are thus considered to be non-pathogenic. Such non-pathogenic F. oxysporum isolates are cosmopolitans and can be found in soil, water and plant residues (Fourie et al., 2011). In addition, non-pathogenic F. oxysporum is a common non-harmful endophyte of plants (Malcolm et al., 2013; Demers et al., 2015). The genetic diversity among non- pathogenic F. oxysporum appears to be much larger compared to the pathogenic members (Lori et al., 2004). Although the term non-pathogenic is well established, care should be taken as it concerns a negative definition. Only by performing pathogenicity tests on a very large collection of plants, a F. oxysporum isolate can truly be called non-pathogenic (Alabouvette et al., 2009).

The observation that soils which naturally restrict Fusarium wilt, better known as a disease suppressive soils (see paragraph 1.3.3.4), harbor large populations of non-pathogenic F. oxysporum triggered the interest in non-pathogenic F. oxysporum as biocontrol agent (Alabouvette et al., 2009; Louvet et al., 1976). Since then, several non-pathogenic F. oxysporum strains have been isolated and studied for biocontrol protection to Fusarium wilt, nematodes, weevils and other plant pathogens such as Pythium ultimum, Phytophthora capsici and Verticillium dahliae (Alabouvette et al., 2009; Benhamou et al., 2002; Mandeel

Chapter 1 and Baker, 1991; Pantelides et al., 2009; Paparu et al., 2006; Postma and Rattinck, 1991; Silvar et al., 2009; Waweru et al., 2014;). Strain Fo47, isolated from the Fusarium wilt suppressive soil of Chateaurenard in France, is the most studied protective strain (Alabouvette et al., 2009). Many other protective non-pathogenic F.oxysporum have been isolated and studied. Non-pathogenic F. oxysporum have a variable protective ability based on different working mechanisms and some strains can even worsen the disease, as has been observed in a greenhouse assay in the banana – Foc pathosystem (Forsyth et al., 2006). Three main working mechanisms have been proposed for protective F. oxysporum strains: competition for nutrients, competition for infection spaces and induction of systemic resistance (Larkin and Fravel, 1999; Mandeel and Baker, 1991). It has been proposed that some strains owe their protective ability to the interaction with ectosymbiotic bacteria (Minerdi et al., 2008). Those bacteria are potentially involved in the expression of some genes of F. oxysporum associated with pathogenicity or protection (Minerdi et al., 2008). Although many strains are effective in greenhouse tests, results obtained on the field are frequently not satisfactory. Generally, a sufficient high ratio of the biocontrol agent to the pathogen is necessary for effective control (for Fo47 a ratio 1/100 pathogen/protective strain is needed for effective control). Protective F. oxysporum strain may suffer difficulties to establish in a highly competitive and variable environment (Alabouvette et al., 2009).

1.3.2 Historical impact of Fusarium wilt on banana

Fusarium wilt on banana is also known as Panama disease, called after the tremendous impact of the fungus on the banana industry in Central America in the first half of the 20th century. At the turn of the 20th century, banana turned from a crop of small scaled farms into a commodity. Big export oriented plantations arose in Costa Rica, Panama, Honduras and Guatemala as vast enclaves under foreign control. Together with the rise of the plantations, problems with Fusarium wilt emerged. Foc was described for the first time in 1876 in Australia and in 1890, it was observed in the banana plantations in Panama and Costa Rica (Ordonez et al., 2015). Gros Michel, the cultivar grown in the plantations, is susceptible to Foc race 1. Although scientists of the United Fruit Company, one of the major banana companies, had already in 1922 found two giant Cavendish cultivars which were resistant to Foc race 1, they only became dominant on export plantations in the 1960’s. Cavendish cultivars were experienced as less tasty and high-demanding in post-harvest as they are more easily bruised. Instead of a cultivar shift, companies moved to disease free areas, a maneuver which was possible because of the powerful position of the heavily vertically integrated companies. This technique was common practice until free areas were no longer

General Introduction available. Other techniques, such as flood fallowing, appeared economically unviable. Heavily dropped profit margins, together with labour problems and reduced support from local governments, finally induced the shift to the resistant cultivars. The cultivar change occurred together with the disintegration of the highly vertically integrated companies as they withdrew from direct production (Moberg and Striffler, 2003).

Nowadays, Foc race 4 (TR4), which affects the commonly grown Cavendish cultivars, is a new threat for the banana industry. TR4 has not yet reached Central America, but impacts heavily the plantations on the Philippines, the main exporter on Asian markets. Breeding projects are investing massively on the creation of a resistant cultivar with good agronomic and consumer properties (Ploetz, 2015; Li et al., 2015). Resistant breeds have been developed (several FHIA hybrids, cv. 925 of CIRAD, Giant Cavendish tissue-culture variants (GCTCV)), but so far, consumer acceptance seems limited (Dorel et al., 2016; Huang et al., 2005). During the Foc race 1 epidemic, little information was available on the epidemiology of the pathogen, causing an easy and wide dissemination of the pathogen. Nowadays, Foc is better known, which allows to take effective measures to limit the spread of TR4. Despite those measures, Foc TR4 seems to spread steadily over the banana growing areas (Butler, 2013; Garcia et al., 2014; Ordonez et al., 2016). It is hard to predict how the banana industry will react on the arrival of TR4 and if it will recover like with the Foc race 1 epidemic. From an optimistic viewpoint, TR4 can act as a driver to system change, towards systems that make greater use of ecological engineering, benefiting from the self-organization of nature and towards investments in an integrated approach to manage diseases and pests (Blouin et al., 2013; Daniells and Lindsay, 2016).

Although the impact on export plantations is better reported, Fusarium wilt also affects small holder banana farmers, who are responsible for the largest share of total banana production. They grow many different cultivars, also cultivars susceptible to Foc race 1. Although no longer of interest to big companies, Foc race 1 stays a major pathogen in banana production. For example, Gros Michel and Silk cultivars, which are susceptible to Foc race 1, are more important in Africa than Cavendish (Lescot, personal communication). Cavendish cultivars are not exclusively grown in export oriented plantations. Worldwide about 30% of all bananas grown for local markets are Cavendish cultivars, making TR4 also a major threat for smallholders (Lescot, 2015).

Chapter 1

1.3.3 Management of Fusarium wilt

1.3.3.1 Preventive measures

To avoid problems with Fusarium wilt, it is of primary importance to prevent the introduction of the pathogen on the field. Once infested, a field stays infested for several decades. Therefore, preventive measures are paramount to safeguard non-infested field from Foc. Limitation of the dissimination of the pathogen is built on two aspects: the containment of the pathogen within the infected areas and the exclusion of the pathogen from disease free areas. Exclusion can be effectuated by good farming practices, such as the use of disease free planting material obtained from tissue culture, no exchange of tools among farms and planting a ground cover to limit erosion (Ploetz, 2015; Vezina, 2015). Quarantine measures, such as food baths and fencing of affected areas, are essential to restrict the pathogen on the affected farms (Ploetz, 2015). Early recognition of the disease and diagnosis are vital for timely containment. From all counties with TR4, Australia has the most rigid biosecurity program. The quarantine measures which have been taken in Australia to constrain Foc TR4 appear not completely waterproof. Recently, TR4 migrated from the Northern territory to Northern Queensland (Ordonez et al., 2015). Although TR4 seems to spread steadily over the world, quarantine measures are capital to slow down the dissemination and restrict the span of affected areas (Butler, 2013; Garcia et al., 2014; Ordonez et al., 2016).

1.3.3.2 Chemical, physical and biological control

Affected plants cannot be cured by chemical application due to the hidden nature of the pathogen. Fumigation has been practiced to eliminate the pathogen in soil. Besides being uneconomical and harmful to the environment, the field appears to be fast re-invaded by Foc (Ploetz, 2015). Many biocontrol agents have been tested to control Fusarium wilt of banana. Despite promising results in greenhouse assays, few have been tested on the field and those tested did not seem to reduce the disease incidence significantly (Belgrove et al., 2011; Ploetz, 2015). Difficulties to establish in soils with different microbial and chemical composition could explain the poor protection in the field by biocontrol microorganisms (Belgrove et al., 2011). Physical control measures such as rice hull burning, solarization and flood fallowing have the objective to eliminate the Foc population in soil but appear not effective (Hermanto et al., 2012; Ploetz, 2015; Stover, 1962).

General Introduction

1.3.3.3 Resistant cultivars

The use of resistant cultivars is often considered to be the most effective method to grow bananas in affected soils (Ploetz, 2015). The cultivar choice depends on the potential severity of Fusarium wilt, the disease and pest pressure of other organisms, agronomic properties and the market preferences. Cultivars with increased resistance to Foc can be obtained by conventional breeding, mutation breeding and genetic engineering (GMO). Several breeding programs exist, of which the Fundácion Hondureña de Investigación Agrícola (FHIA) is the oldest. Several hybrids resistant to Foc race 1 and 4 have been developed, but with inferior agronomic, post-harvest or taste properties in comparison to the traditional cultivars (Ploetz, 2015). Traditional breeding faces great challenges due to the polypoid nature of the crop, long generation time, large plant sizes, genetic abnormalities in many parental lines, the lack of information on the genetic basis of resistance to Foc in banana, and the need for a sterile and parthenocarpic final product (Li et al., 2015; Ploetz, 2015). The production of somaclonal mutants is a second strategy, which has resulted in the development of the so-called Giant Cavendish Tissue Culture Variants (GCTCV) in Taiwan. These Cavendish mutants have an increased level of resistance to Foc race 4, but need to be replaced after one or two growth cycles in TR4-infested soil (Ploetz, 2015). Finally, the production of GMO’s may be theoretically promising, but so far, no GMO has made it to the market (Ploetz, 2015).

1.3.3.4 Disease suppressive soils

The soil can have a major influence on the occurrence of disease. In some soils little or no disease develops under conditions that are seemingly favorable for disease development. Those are best known as disease suppressive soils (Baker and Cook, 1974; Kinkel et al., 2011). On the contrary, in conducive soils, even a normally incompatible interaction between a pathogen and resistant host can result in disease (Furtado et al., 2009; Peng et al., 1999). It has been observed that Cavendish cultivars, which are normally resistant to Foc race 1, can be affected by Foc race 1 when grown in waterlogged soils or under drought conditions (Peng et al., 1999). The capacity of a soil to suppress a disease ranges along a continuum from highly suppressive to very conducive (Alabouvette and Steinberg, 2006). The enhancement of the disease suppressive capacity of the soil is a desirable trait to control Fusarium wilt diseases.

Chapter 1

A more diverse and active microbial community will lead to general disease suppression by fungistasis. The germination and growth of the pathogen is inhibited through the highly competitive microbial community (Alabouvette and Steinberg, 2006). In specific suppression, the presence of a specific organism can antagonize the pathogen or strengthen the plant (Alabouvette and Steinberg, 2006). Despite the conceptual separation of general and specific suppression, scientists acknowledge that natural suppressiveness is the outcome of general and specific organisms in association with the abiotic conditions (Alabouvette and Steinberg, 2006, Cook and Baker, 1983; Garrett, 1970).

Examples of soils suppressive to Fusarium wilt on diverse crops are ample, including on banana (Alabouvette, 1986, 1999; Domínguez et al., 2001; Larkin et al., 1993; Nel et al., 2006; Peng et al., 1999; Scher and Baker, 1980). In banana, the length of time that production of susceptible cultivars in presence of the pathogen can be maintained is the principal indication of the suppressive capacity (Ploetz, 2015; Stover, 1962). The main abiotic soil properties that have been correlated with Fusarium wilt suppression are the presence of clay minerals such as montmorillonite and illite, a higher pH and a higher structural aggregate stability (Alabouvette, 1999; Amir and Alabouvette, 1993; Domínguez et al., 2001). Besides a more active and versatile general soil microbial community, non-pathogenic F. oxysporum and fluorescent Pseudomonas are the microbes most often associated with suppressive soils (Alabouvette, 1999; Shen et al., 2015). Their suppressive capacity is largely attributed to competition for carbon and iron, respectively (Alabouvette, 1999). However, to our knowledge, no successful establishment of soil disease suppression has been achieved upon manipulation of an abiotic soil factor or introducing non-pathogenic F. oxysporum and Pseudomonas as biological control agents. Managing one single factor is probably insufficient to induce soil disease suppression since this is the outcome of the interaction of multiple factors (Alabouvette, 1999).

Adapting cultural practices is an alternative strategy to impact the disease suppressive capacity of a soil (Alabouvette and Steinberg, 2006). A lot of research has been done on the effect of organic amendments on Fusarium wilt suppression. Amendment with compost has been shown to be very effective to control Fusarium wilt in several crops, however there is little information on compost admentments in the banana – Foc pathosystem (Ploetz, 2015; Serra-Wittling et al., 1996; Termorshuizen et al., 2006). In a recent field study, increased disease suppression was obtained after amending banana plantlets with a compost which has been inoculated with the potential biocontrol agent Bacillus amyloliquefaciens (Xue et al., 2015). Another possible cultural practice that can influence disease suppression can be the alternation of crops, both in time (crop rotation) or in space (intercropping). Stover (1962)

General Introduction remarked that mixed plantings of banana developed more moderate losses to Fusarium wilt in comparison to monocultures. In China, several farmers remarked that rotation with Chinese leek (Allium tuberosum) reduced Fusarium wilt in the following banana cropping cycle (Huang et al., 2012). This effect was attributed to the antifungal effect of its root exudates and volatiles (Zhang et al., 2013a). In other crops, Fusarium wilt could be suppressed by companion cropping, for example the combination watermelon with rice or wheat reduced Fusarium wilt in watermelon in pot experiments (Hao et al., 2010; Xu et al., 2015). On the contrary, in a single case, the induction of Fusarium wilt suppression was obtained by continuous monoculture. Continuous growth of a moderately susceptible musk melon rendered the soil suppressive (Hopkins et al., 1987; Larkin et al., 2003). No cases are known of induction of Fusarium wilt suppressiveness by continuous monoculture of banana (Ploetz, 2015).

1.4 The soil ecosystem: a complex of interactions

The soil is besides the matrix for plants to establish, one of the most complex ecosystems on earth. Its multiple and diverse habitats accommodate a diverse range of organisms, which can be split into hierarchical groups. Microorganisms, such as bacteria and fungi make up the smallest scale. They are followed by their predators, such as nematodes, which on their turn are followed by the so-called ecosystem engineers: plants, earthworms, termites, ants etc. Those ecosystem engineers play a key role in shaping habitats for other organisms and controlling their actions through physical and biochemical interactions. Plants have the largest influence in the direct environment of their roots, called the rhizosphere. Likewise, earthworms have the biggest impact in the drilosphere, termites in the termitosphere and ants in the myrmecosphere (Lavelle et al., 2016). A stable soil ecosystem is the outcome of complex interactions between the different members. In this section, we will introduce the interactions between plants and microorganisms in soil. Figure 1.5 shows the main modes of interaction between plants and microorganisms. However not covered, it is important to keep in mind that other actors, such as rainworms and termites, and the environment will impact and alter the plant-microbe interactions as well.

It is well established that plants affect the soil microbial community (Broeckling et al., 2008; Garbeva, 2004; Kinkel et al., 2011). They can influence the microbial community by modulating the physical environment. However, the largest influence of plants on soil microorganism is attributed to their root exudates. Plants excrete up to 40% of their photosynthetic constructions to the environment via their roots. Those exudates comprise of

Chapter 1 compounds of small molecular weight, such as sugars, amino acids, ions, organic acids, secondary metabolites and compounds of high molecular weight such as mucilage (polysaccharides) and proteins. The latter forms the largest proportion of the exudate by mass, while the first accounts for the majority of the diversity in root exudates (Bais et al., 2006). Excretion of root exudates can occur either passively, by leaching, or via an active process (Bais et al., 2006). The composition of root exudates is dependent on factors such as plant species, age, cultivar, physiological state and stressors (Bais et al., 2006; Doornbos et al., 2012).

Root exudates can stimulate the growth of microorganisms by delivering an easily accessible carbon substrate. Together with other rhizodepositions, such as root debris and border cells, they form the major source of organic C in the soil (Chaparro et al., 2012). The excretion of specific signaling molecules is crucial to establish symbiotic interactions, such as with nitrogen-fixating bacteria and vesicular arbuscular mycorrhizas. Root exudation can have a negative impact on microbes by containing fungicidal, antibiotic or quorum sensing inhibiting compounds (Bais et al., 2006; Raaijmakers et al., 2009). Some plants discourage microbial growth by decreasing the soil pH in the rhizosphere (Chaparro et al., 2012). On their turn, microorganisms will influence the plant community by altering the fitness of individual plants. Endophytic associations can lead to increased tolerance to biotic and abiotic stresses. Other associations such as with the aforementioned nitrogen-fixing rhizobia or ecto-and endomycorrhizal fungi, can influence the nutrient availability for plants. Some plant growth promoting rhizobacteria (PGPR) can improve growth and health directly, for instance by excretion of phytohormones and their precursors (Bais et al., 2006; Chaparro et al., 2012; Lugtenberg and Kamivola, 2009). The fitness of plants can be reduced by plant pathogens and production of microbial phytotoxic compounds. Microorganisms can alter on their turn the plant root exudates and subsequently, impact the microbial community (Chaparro et al., 2012; Steinkellner et al., 2008).

Besides interactions between plants and microorganisms, mutual interactions also have a considerable impact on the formation of the soil community. Plants mutually compete for resources such as nutrients, light and water. Root exudates can affect proximate plants negatively by phytotoxins, also known as allelopathic compounds, and the excretion of signal molecules to establish a parasitic association or to attract pathogens of the surrounding plants (Inderjit and Weiner, 2001; Mangla and Callaway, 2008). The alteration of nutrient availability, by phytosiderophores (Fe) or organic acids (P) can be both negative and positive for neighbouring plants (Chaparro et al., 2012). After attack by herbivorous insects, some plants can induce resistance against those herbivores in their neighbours via the root

General Introduction exudates (Bais et al., 2006; Inderjit and Weiner, 2001). In addition, plants influence and communicate with each other via an extensive network of mycorrhizal fungi that interlinks plants (Simard and Durall, 2004; Van Der Heijden and Horton, 2009).

Also microbe-microbe interactions have a community shaping impact. Competition for nutrients, antagonism and hyperparasitism are the main types of interaction (Raaijmakers et al., 2009; Kinkel et al., 2011). Antagonistic action includes the production of secondary metabolites with antibiotic function, the excretion of lytic enzymes and effectors. Several microbes have developed mechanisms to escape or wreck antagonistic attacks of fellow microbes (Raaijmakers et al., 2009).

Figure 1.5: The main interactions between plants and soil micro-organisms. Own design based on information from Bais et al. (2006), Raaijmakers et al. (2009), Chaparro et al., (2012), Inderjit and Weiner, 2001.

The interactions described above lead to a continuous changing environment, pursuing the highest energy capture efficiency and ecosystem stability (Lavelle et al., 2016). The microbial community will change the plant community and the changed plant community will on its turn influence the microbial community. The stability within the soil ecosystem is built on this continuous interplay.

Chapter 1

In agriculture, the plant community is managed and imposed. This implicates the disturbance of the continuous interplay between the plants and the other organisms towards the most energetic effective system. Problems caused by soil-borne plant pathogens in agricultural systems can often be situated in this frame. In stable natural systems, serious pest and disease epidemics are rare, since pest and disease populations are brought automatically back into balance by natural control (Gordon and Martyn, 1997; Jorgensen and Nielsen, 1997; Van der Veken, 2010). Although disturbance of the natural ecosystem is inherent to agriculture, cropping systems can be designed in such manner that they make optimal use of beneficial ecological interactions. Like illustrated above, the interactions are multiple and much more knowledge is required to effectively use them in the design of ecological efficient cropping systems.

1.5 Problem statement and thesis outline

Management of Fusarium wilt is challenging because of the soil-borne character and long term survival of the spores and the sterility of most modern banana cultivars. It becomes apparent that no one-sided solution will be sufficient to cope with this disease. The use of resistant cultivars is put forward as the most effective way to grow bananas on infested soils. However, Foc seems to have an adaptive capacity, as has been observed with the rise of TR4, and breeding programs tend to run behind the problem. Until now, no biological control has proven efficacy on the field, which is probably due to poor establishment of the biocontrol agent under field conditions. Several cultural measures appear to contribute to the disease control, but are individually often insufficient to reach the desired level of control. A better understanding of the epidemiology, survival, virulence of the pathogen and interaction with its biotic and abiotic environment is indispensable to attain effective control of Fusarium wilt in banana.

As mentioned in paragraph 1.3.3.4, increasing the disease suppressive capacity of the soil would be a potential option to manage Fusarium wilt in banana. Since disease suppression is mainly microbiological in origin, it is the product of the soil ecological interaction as illustrated in paragraph 1.4. This PhD project started from a remarkable farmer’s observation. A Brazilian farmer noted a large variability in the disease severity of Fusarium wilt in his highly diverse banana cropping system. Our starting research question (RQ) was as follows:

General Introduction

RQ 1 (chapter 2): Can the observed differences in Fusarium wilt in the field be attributed to differences in soil disease suppressiveness? If yes, which soil abiotic properties, plant community or microbial factors are correlated with Fusarium wilt suppressiveness in soil?

In chapter 2, it was verified if the observed variability in the field could be attributed to different levels of disease suppression in soil. Subsequently, to identify potential factors influencing the disease suppressive capacity of the soil, locations with lower and higher level of disease suppression were compared for their soil abiotic properties, the plant community and the microbial soil community. Based on this study, three additional research questions were formulated.

RQ 2 (chapter 3): How does Foc relate to the local non-pathogenic F. oxysporum population? RQ 3 (chapter 4): Do graminoids contribute to pathogen inoculum build-up as symptomless carrier? RQ 4 (chapter 5): Can banana cultivars with different level of resistance influence disease suppression in soil? Subquestion: How do different cultivars influence the Foc and non-pathogenic F. oxysporum population?

Fourie et al. (2011) stated in their review about the current status of Foc within the FOSC that “The origins and nature of genetic variation in Foc are important subjects for future study. Future research should further focus on in-depth comparisons between closely related formae speciales and nonpathogens of F. oxysporum, rather than focusing strictly on pathogens of agricultural crops.” Based on this research need, F. oxysporum isolates collected on the farm were subjected to a molecular population study in chapter 3. Foc isolates with a distinct and a similar EF-1α/IGS sequence type than non-pathogenic isolates were found. To explore if Foc isolates showing similarity with the non-pathogenic population gained pathogenicity through individual evolution or by transfer from pathogenicity genes from the other Foc isolates present on the farm, the presence and sequence of putative effector genes were studied.

In the most disease conducive locations of the farm graminoids were abundant. We hypothesized that graminoids may contribute to the increase of pathogen inoculum as symptomless carriers of Foc. However, explorative field sampling of graminoids on the field resulted exclusively in non-pathogenic F. oxysporum isolates. Therefore, in chapter 4, we investigated whether Foc and non-pathogenic F. oxysporum isolates differ in their capacity or

Chapter 1 competitiveness to colonize the graminoids Cyperus iria and Brachiaria decumbens. Additionally, the growth profile of the different isolates was studied to investigate the relation between growth and the competitive ability to colonize the roots.

Besides the influence of graminoids, field observations suggested an important role for different banana cultivars in the outcome of the Fusarium wilt. In chapter 5, we studied the interaction of Foc and non-pathogenic F. oxysporum with four banana cultivars with different level of Foc race 1 resistance. Both the root colonization and the soil population of Foc and non-pathogenic F. oxysporum isolates were monitored for the different cultivars. Root exudates are an important interface between plants and microbes. Therefore, the root exudates of the different cultivars with and without Foc inoculation were collected and compared for their impact on the germination of microconidia. In addition, the sugar and amino acid content of the root exudates was analysed.

Finally, in chapter 6, the general conclusions from the work are summarized, discussed and future perspectives are articulated. Figure 1.6 illustrates schematically the structure of this thesis.

Figure 1.6: Overview scheme of the different chapters of this PhD thesis.

Chapter 2

Disease suppressiveness to Fusarium wilt of banana in an agroforestry system: Influence of soil characteristics and plant community

Authors Pauline Deltoura,b, Soraya C. Françaa, Olinto Liparini Pereirab, Irene Cardosoc, Stefaan De Neved, Jane Debodee, Monica Höftea

Affiliations aLaboratory of Phytopathology, Department of Crop Protection, Faculty of Bioscience Engineering, Ghent University, Coupure Links 653, 9000, Ghent, Belgium bLaboratório de Micologia a Etiologia de Doenças Fúngicas de Plantas, Departamento de Fitopatologia, Universidade Federal de Viçosa, Avenida Peter Henry Rolfs s/n, 36570-900, Viçosa, MG, Brazil cDepartamento de Solos, Universidade Federal de Viçosa, Avenida Peter Henry Rolfs s/n, 36570-000, Viçosa, MG, Brazil dDepartment of Soil management, Faculty of Bioscience Engineering, Ghent University, Coupure Links 653, 9000, Ghent, Belgium ePlant Sciences Unit – Crop Protection, Institute for Agricultural and Fisheries Research (ILVO), Burg. van Gansberghelaan 106, 9820, Merelbeke, Belgium

Published in Agriculture, Ecosystems and Environment, February 2017, Volume 239, 173–181.

Chapter 2

Abstract

One of the most destructive diseases of banana is Fusarium wilt or Panama disease, caused by the soilborne fungus Fusarium oxysporum f. sp. cubense (Foc). Foc is widespread in almost all banana-growing areas and cannot be effectively controlled by chemical or biological measures. Fusarium wilt could potentially be managed by the promotion of soil disease suppressiveness, but little is known how soils attain a higher level of disease suppression and how crop management can impact on this. Banana cultivar (cv.) Maçã, a cultivar highly susceptible to Foc race 1, was grown on a farm managed as agroforestry system in Pedra Dourada, Brazil, where Foc race 1 is present in soil. At some locations on the farm banana plants of cv. Maçã stayed productive, while on others it succumbed rapidly. We hypothesized that the differences in disease severity on the farm could be attributed to different levels of soil disease suppressiveness. In this study, we assessed the level of disease suppression of the different locations and elucidated potential factors that could promote disease suppression in soil. Patches with confirmed presence of Foc race 1 were sampled and tested for Fusarium wilt suppression in greenhouse assays. The plant community composition, soil abiotic properties and soil microbial community of the different locations were compared. Locations with a higher level of disease suppression were characterized by a low density of the susceptible cv. Maçã, a high diversity of other banana cultivars, a higher clay content, higher pH and lower soil cover by graminoids. Banana cv. Ouro was only present in the three most suppressive patches. The results of this study suggest that in soils with favorable abiotic properties, a good plant arrangement, in which cv. Maçã is grown in mixed stands with other banana cultivars, can help to promote Fusarium wilt suppression.

Disease suppressiveness to Fusarium wilt in an agroforestry system

2.1 Introduction

In tropical regions, banana is of high socioeconomic importance, since it is a cash crop in export-oriented plantations, and a food- and cash crop on smallholder farms. Banana is a robust crop with high returns on input, but production of certain cultivars is often restrained by Panama disease (Fusarium wilt), a wilt disease caused by the soil-borne fungus Fusarium oxysporum f.sp. cubense (Foc) (de Oliveira e Silva et al., 2001). Since Foc can produce long-term survival structures, the chlamydospores, or survives as a saprophyte on non-host plants, it is virtually impossible to get rid of the pathogen once the soil is infested. Foc is disseminated via contaminated soil on wheels and shoes, contaminated tools, runoff and irrigation water, latently infected planting material and insects (Meldrum et al., 2013; Ploetz, 2015). So far, Foc levels in soil cannot reliably be monitored as the pathogen is not morphologically distinguishable from non-pathogenic F. oxysporum, and molecular detection methods are still in development (Dita et al., 2010; Peng et al., 2014).

Based on pathogenicity to specific cultivars, Foc has been subdivided into four races, three of which are of economic importance (i. e. race 1, 2 and 4). Cultivars of the groups Gros Michel (AAA) and Silk (AAB) are susceptible to race 1, while cooking bananas, such as Silver Bluggoe (ABB), are susceptible to race 2. Race 4 can affect cultivars of the Cavendish subgroup (AAA) and all cultivars susceptible to race 1 and 2. Race 4 is subdivided in a subtropical (ST4) and tropical (TR4) race 4 depending on the climatic condition in which it can affect Cavendish cultivars. In contrast to Foc race 1, which has been reported in almost all banana growing regions (Ploetz, 1990), race 4 is mainly limited to South-East-Asia and Australia, but has recently been observed in Jordan, Pakistan, Lebanon, Mozambique and Oman (Dita et al., 2010; Ordonez et al., 2015). In export-oriented plantations the problem with Foc race 1, which devastated great areas in the first half of the 20th century, was solved by a shift from the susceptible cultivar Gros Michel to cultivars of the resistant subgroup Cavendish. However, Foc race 1 remains an important challenge in countries such as Brazil, where bananas are mainly grown for local consumption (Borges and da Silva Souza, 2004; Ploetz, 1990).

In Brazil, 98.8% of the total banana production is sold on local markets (Borges and da Silva Souza, 2004; Lescot, 2015), which are dominated by AAB cultivars, which differ widely in their susceptibility to Foc race 1 (Table 2.1). The area planted with cv. Maçã has decreased due to its high susceptibility to Foc race 1, but this cultivar is very profitable because it is preferred by consumers (Moreira, 1999; personal communication Cândido da Silva). In

Chapter 2 addition, cultivars of the Cavendish subgroup and many cultivars of minor importance, which are in general resistant to Foc race 1, are cultivated (Table 2.1, Borges et al., 2009).

Table 2.1: The major banana cultivars grown in Brazil and their susceptibility to Foc race 1 (Borges et al., 2009)

Genomic Susceptibility Subgroup Cultivar group to Foc race 1 AAB Pome Prata MS Pome Pacovan MS

Pome Prata anã MS

Silk Maçã VS

Mysore Mysore R

Plantain Terra R

Plantain Terra d'Angola R

AAA Cavendish Nanica R Cavendish Nanicão R

Cavendish Grande Naine R

AA Sucrier Ouro R ABB Figo Figo Cinza R Figo Figo Vermelho R

VS: Very susceptible MS: Moderately susceptible R: Resistant

In susceptible cultivars, Foc is difficult to control since fungicide application is ineffective and uneconomical, and biological control has no proven efficacy in the field (Ploetz, 2015). However, some soils have the potential to suppress Fusarium wilt in spite of the presence of a virulent pathogen (Abadie et al., 1997; Baker and Cook, 1974). Those soils are found worldwide for a number of crops, including banana (Alabouvette, 1986; Alabouvette, 1999; Domínguez et al., 2001; Larkin et al., 1993; Nel et al., 2006; Peng et al., 1999; Scher and Baker, 1980). The capacity of a soil to suppress diseases is ranging along a continuum from highly conducive to suppressive (Amir and Alabouvette, 1993; Janvier et al., 2007). Soil microorganisms play key roles in suppressing soil borne diseases mainly via antagonism or competition. Plants, soil type and microbial interactions are believed to be the main driving forces of the microbial community in soil (Broeckling et al., 2008; Grayston et al., 1998, Wieland et al., 2001) and can thus influence its disease suppressive potential (Garbeva et al., 2004; Janvier et al., 2007, Kinkel et al., 2011).

Disease suppressiveness to Fusarium wilt in an agroforestry system

In Pedra Dourada, Brazil, a farmer observed that banana cv. Maçã (Subgroup Silk, AAB) succumbed to Fusarium wilt on certain locations of his farm, while on others it stayed productive. The farm is managed as an agroforestry system with very diverse plant arrangements. We hypothesized that the variation of disease expression on the farm was due to different degrees of disease suppressiveness in soil. The objective of this study was to assess the disease suppression in the different locations and to identify essential soil and plant community factors that contribute to disease suppressiveness. Knowledge about these factors can lead to optimized crop management strategies that enhance the disease suppressive potential of soils.

2.2 Material and methods

2.2.1 Studied site

The studied farm (12 ha) is located in Pedra Dourada, a municipality of the region Zona da Mata, Minas Gerais, Brasil (20°50’ S, 42°07’ W) in the Atlantic rainforest. The region is dominated by family farmers, some of whom are actively involved in the development and implementation of agroforestry systems through participatory research projects with the support of the Centre for Alternative Technologies of Zona da Mata (NGO), the Federal University of Viçosa and organizations of family farmers (Cardoso et al., 2001).

Since 1995, the farm has been managed based on agroecological principles, without the use of synthetic fertilizers or pesticides. The farm consists of two parts with different land use history. Part A was planted with full sun coffee intercropped with few scattered banana plants before it was converted into an agroforestry system in 1995. The main crops are coffee and banana (cv. Maçã, Prata, Ouro, Nanica, Terra), intercropped with fruit and native trees. There is no use of rows, patterns and no fixed planting distances. The banana plants are only desuckered to obtain new planting material. Part B was until 2004 a pasture with Bracharia brizantha. In 2004, 2000 suckers of cv. Maçã were bought from a neighboring farmer and planted on the farm. One part was planted on available spots in part A, the rest was planted in area B. In area B, planting holes were dug in the pasture, with no fixed planting distances, as the area was intended to be intercropped in the future. Three years

Chapter 2 later, all banana cv. Maçã planted in part B had succumbed due to Fusarium wilt, while in part A survival was variable. Of the plants that succumbed to the disease, still mats with small tiny suckers remained on the field. At the moment of the study, area B was in conversion to an agroforestry system.

Twenty plants of banana cv. Maçã were sampled (Figure 2.1): 18 plants in area A, where the survival was variable (PDa1-PDa18) and two plants in area B, where only small infected mats were left (PDb1, PDb2). With the participation of the farmer, plants were selected and evaluated for bunch production (0: no bunch production, 0.5: production of bunches with 5 hands or less , 1: production of bunches with more than 5 hands) and length of highest sucker (Adapted from Carlier et al., 2003).

Figure 2.1: The studied farm with the distribution of the sampled banana plants in area A and area B. The water divides the farm in a northern and southern faced slope. The use of the white areas is diverse, such as the house, vegetable garden, a pasture.

Disease suppressiveness to Fusarium wilt in an agroforestry system

2.2.2 Isolation, pathogenicity and race identification of Foc

The presence of Foc was checked in all 20 plants sampled. Plant pieces of the base of the pseudostems with vascular discoloration were surface sterilized in NaOCl and plated on Potato Dextrose agar (PDA) with 10 mg l-1 rifampicin (for 1 l PDA: an infusion of 200 g potatos, 20 g dextrose, 15 g agar, complemented with distilled water). Isolation from plant tissue was performed, since F. oxysporum is ubiquitous in soil and pathogenic F. oxysporum cannot be distinguished morphologically from non-pathogenic individuals. The isolates that were morphologically identified as F. oxysporum, were kept as single spore cultures on dried filter paper, and later used for molecular analysis and pathogenicity testing. DNA was extracted with the Wizard Genomic DNA Purification kit (Promega). PCR amplifications of the translation elongation factor (EF-1α) fragment (650 bp) and the intergenic spacer (IGS; 2200 bp) were performed using the primers EF1 and EF2 (O’Donnell et al., 1998), and the primers iNL11 and iCNS1 (O’Donnell et al., 2009), respectively. After purification (E.Z.N.A. cycle pure kit, Omega, VS) PCR products were sent for sequencing (LGC Genomics, Berlin, Germany). The IGS regions were sequenced with two additional primers NLa and CNSa (O’Donnell et al., 2009). The sequences were blasted in Fusarium-ID for species and multilocus type identification (Geiser et al., 2004; O’Donnell et al., 2009).

The isolates were tested for pathogenicity on cv. Maçã (susceptible to races 1 and 4) and cv. Grande Naine (susceptible to race 4), using pathogen-free plantlets from tissue culture (Multiplanta Tecnologia Vegetal, Andradas, MG). Four plantlets per isolate were dipped for 2 h in a conidiospore suspension (106 spores ml-1) prepared from 10-day-old cultures on PDA. As control, four plantlets were dipped in sterile distilled water. After inoculation, plants were grown in 500 ml pots filled with substrate (Tropstrato HT, vida verde, Mogi Mirim, SP) in a greenhouse (18-28°C average night-day temperature; 12:12 light regime). After 7 weeks, the plants were evaluated for the absence or presence of external disease symptoms: yellowing of the leaves, splitting of the pseudostem and stunted growth. Internal symptoms were assessed by using the Rhizome Discoloration Score (RDS), based on the discolored area of the stellar region (0: no discoloration of the stellar region; 1: 1-25%; 2: 26-50%; 3: 51- 75%; 4: 76-100%) (Adapted from Mak et al., 2004). A disease index was calculated: Disease index = ∑(RDSi X number of plants with RDSi) / (maximal RDS X total number of plants). The pathogen was reisolated from root, rhizome and leaf pieces to confirm Koch’s postulate. Isolates that could successfully be reisolated from plants showing external symptoms and vascular discoloration were considered pathogenic.

Chapter 2

2.2.3 Soil sampling and assessment of disease suppression

Soil samples were collected around the plants infected with Foc race 1 that had an identical multilocus type (6 plants in area A and 2 plants in area B). In June 2013, four soil samples from 0 to 20 cm depth were collected at 50 cm distance of each selected plant of banana ‘Maçã’. The samples were bulked and a subsample was used for physical-chemical and microbial analyses.

A chlamydospore-rich inoculum was prepared according to the method of Smith and Snyder (1971). A fraction of the collected soil samples was dried and autoclaved for 2 consecutive days during 1h and subsequently inoculated with a conidiospore suspension of FoxPDa15 (race 1, 107 spores g-1 soil) and incubated for 3 weeks at room temperature for development of chlamydospores. Final inoculum levels were in the range of 105 cfu g-1 soil.

To test disease suppression, collected soil samples were inoculated with the chlamydospore inoculum, reaching two final concentrations: 5.103 and 5.104 cfu g-1 soil. Tissue culture- derived plantlets of banana cv. Maçã (Multiplanta Tecnologia Vegetal, Andradas, MG) were planted in the infested soils in pots of 500 ml and grown for 3 months in the greenhouse (18- 28°C average night-day temperature; 12:12 light regime). Each concentration was applied in three replicates per soil sample. Non-inoculated soil samples were also evaluated for natural infestation level and consisted of five replicates per soil sample. Pots were weekly supplied with Hoagland nutrient solution. Disease incidence, based on external symptoms, was evaluated weekly from 46 until 98 days after planting. The relative area under the disease progress curve (RAUDPC) was calculated by dividing the area under the disease progress curve by the maximum area. At the final evaluation, internal symptoms were evaluated by the RDS and a disease index was calculated as described before. Finally, Foc was reisolated to confirm Koch’s postulate.

The level of disease suppression of each patch was determined by grouping the information of the different parameters evaluated in the greenhouse assay, namely the disease index and the RAUDPC, and considering both inoculum concentrations. The different soil samples were ranked from high to low RAUDPC or disease index, with mean values for the tie ranks, and a joined rank was calculated as the mean of all the separate rankings. Statistical differences were tested with a multiple comparison (Kruskal Wallis) on the rank values. Similarly, the scores for bunch production and length were ranked from low to high and a joined rank was calculated as measure for disease suppression in the field. High ranks

Disease suppressiveness to Fusarium wilt in an agroforestry system indicate a higher level of disease suppression. The Spearman’s correlation between the plant performance in the field and the disease suppression as observed in the greenhouse assay was calculated based on the joined ranks.

2.2.4 Characterization of the patches

2.2.4.1 Soil physical-chemical properties

Analyses of soil abiotic properties were performed at the laboratory of Soil management, Ghent University. Soil texture was analyzed with the pipet method of Robinson-Köhn (ISO11277:1998). The pH (KCl) and electric conductivity (EC) were measured with an electrode (ISO 10390:2005 and ISO 11265: 1994, respectively). Total nitrogen was measured by catalytic combustion gas separation coupled with a thermal conductivity detector (VarioMAX CNS analyzer Elementar, Hanau, Germany) and soil organic carbon via the Walkley-Black method. Differences in the physical-chemical properties of the soil samples were visualized by principal component analyses and spearman’s correlation coefficients were used to characterize the relationship between the joined rank of disease suppression as determined in the greenhouse assays, and the soil variables.

2.2.4.2 Vegetation parameters

As the plant community on the farm was very diverse, all perennial plants in a radius of 5 m around the sampled banana plants were counted and mapped. Those comprised of adult trees and shrubs, namely plants that are part of the canopy and older than 2 years, juvenile trees and shrubs, namely the plants that are part of the understory and younger than 2 years, coffee and banana. For every survey patch, we determined the abundance and calculated the Shannon-Weaver diversity (H’) of the adult trees and shrubs, juvenile trees and shrubs, and banana cultivars (H’= -∑ pi ln(pi) with pi the proportion of a tree species or cultivar on the area). The soil cover at the base of the selected plant, which generally consisted of graminoids, was attributed a score according to the area covered: 1: 0%; 2: 0-25%; 3: 25- 50%; 4: 50-75%; 5: 75-100%. The degree of shadow was scored on a fully sunny day at two time points (morning and afternoon). Scores were attributed to the degree of shadow on the ground in a radius of 2 m around the banana plant (1: 0% shadow; 2: 0-25% shadow; 3: 25- 50% shadow; 4: 50-75%; 5: 75-100%) and the mean was used as general score for shadow.

Chapter 2

To visualize differences in the species assemblies of the different patches a principal component analysis was conducted. Spearman’s correlation coefficients were used to characterize the relationship between the joined rank of disease suppression as determined in the greenhouse assays and the plant community. For adult and juvenile tree and shrub species, only the abundance was used as the Shannon-Weaver diversity was highly correlated with the abundance because almost all trees were of different species. The composition of banana plants was integrated in the analysis by the total abundance of banana plants, the proportion of banana cv. ‘Maçã’ and the diversity of different cultivars

(Maçã, Prata, Ouro, Nanica and Terra), calculated as a Shannon-Weaver index (H’= -∑ pi ln(pi) with pi the proportion of cultivar i). Also the soil cover and degree of shadow were taken into account as they are mostly determined by the nearby canopy.

2.2.4.3 Soil microbial community

Populations of culturable bacteria were estimated by dilution plating and counting on Luria- Bertani agar (LB). The density of culturable bacteria is used as an indicator for the size of the total bacterial community. The density of Fusarium spp. in soil was estimated by dilution plating on Komada’s medium, a Fusarium selective medium (Komada, 1975).

The composition and diversity of the dominant bacterial species was analyzed with denaturing gradient gel electrophoresis (DGGE). After DNA extraction from 250 g lyophilized soil per sample (Mo Bio Ultraclean Soil DNA), DNA of the V3 fragment of the bacterial 16S rRNA gene was amplified with a nested PCR. For the first round, the near full length 16S rRNA gene was amplified with the primers pA: agagtttgatcctggctcag and pH’: aaggaggtgatccagccgca (Hall et al., 1999) and a standard 50 μl PCR reaction mixture as described in Debode et al. (2013). The following PCR program was used: 95°C set-up denaturation, 3 cycles of 45 seconds denaturation on 95°C, 2 minutes annealing on 55°C and 1 minute extension on 72°C, followed by 30 cycles of 20 seconds denaturation on 95°C, 1 minute annealing on 55°C and 1 minute extension on 72°C, concluded with a final extension of 7 minutes on 72°C. Second, to obtain the V3 part, the 16S rDNA amplicon (∼1500 bp) from the first PCR was diluted and further amplified with a second PCR using the primers 341F (cctacgggaggcagcag) and 518R (attaccgcggctgctgg), with a GC clamp on the first primer and a touchdown PCR program (Muyzer et al., 1997). The reaction mix was enriched with 10 μl GC-rich. The V3-PCR products were purified using an ethanol precipitation and resuspended in 10 μL water. The DGGE was performed as described in Debode et al. (2013), but with a denaturing gradient from 45% to 65% (v/v) of urea and

Disease suppressiveness to Fusarium wilt in an agroforestry system formamide. The DGGE profiles were compared using band-based comparison using Jaccard-Ward clustering tool, with 15% cut off value, with BIONUMERICS (version 5.1, Applied Math’s, Belgium). The band pattern (i. e. bacterial species) was analysed for relatedness with suppressiveness, physical-chemical soil properties and vegetation parameters by the use of redundancy analysis (RDA). Spearman’s correlation coefficients were calculated between the joined rank of the greenhouse assays data and the microbial indicators.

2.3 Results

2.3.1 Isolation, pathogenicity and race identification of Foc

F. oxysporum could be recovered from the base of the pseudostem of 11 out of the 20 banana plants sampled in the various patches indicated in Figure 2.1. In total, 13 isolates were obtained as two isolates each were obtained from PDa6 and PDa9 (Table 2.2). The thirteen isolates were characterized as F. oxysporum by morphological characteristics and molecularly identified based on the EF-1α sequence. From those, nine were shown to be pathogenic to banana plants. Eight isolates showed pathogenicity on cv. Maçã, but not on cv. Grande Naine, and were characterized as Foc race 1. These isolates had identical EF-1α and IGS sequences, and belong to multilocus type 25, according to the grouping of F. oxysporum proposed by O’Donnell et al. (2009). One plant cv. Grande Naine inoculated with FoxPDa14 showed minor vascular discoloration. As the plant showed no external wilting symptoms and the EF-1α/IGS multilocus type is the same as the other isolates which are all identified as race 1, the observed discoloration is interpreted as an anomaly due to the harsh inoculation method (Pegg and Langdon, 1987). Isolate FoxPDa3 showed clear vascular discoloration and external wilt on cv. Maçã and cv. Grande Naine. FoxPDa3 does not belong to any of the 256 multilocus types described on Fusarium–ID and its IGS and EF-1α sequence differ from TR4 isolate NRRL36114, which belongs to VCG01213/16 (O’Donnell et al., 2009). Soil surrounding the eight plants from which Foc belonging to multilocus type 25 was isolated, was assessed for disease suppression. Patch PDa3 was omitted in the further characterization to avoid variation due to difference in virulence of the pathogen.

Chapter 2

Table 2.2: F. oxysporum isolates obtained from cv. Maçã from the area A (PDa) and area B (PDb) and their pathogenicity on cv. Maçã and cv. Grande Naine as tested in the greenhouse. Accession numbers of the intergenic spacer (IGS) and elongation factor 1α (EF-1α) sequences in GenBank are given.

cv. Maçã cv. Grande Naine Accession number Disease External Disease External Name Patch of index symptoms Pathogenic index symptoms Pathogenic IGS EF-1α isolate isolation (RDS)a b (RDS)a b FoxPDa1 PDa1 100 2/4 + 0 0/4 - KU577995 KU577989 FoxPDa2 PDa2 0 0/4 - 0 0/4 - KU577997 KU577991 FoxPDa3 PDa3 58 2/3 + 44 2/4 + KU577996 KU577990 FoxPDa6 PDa6 0 0/4 - 0 0/4 - KU577998 KU577992 FoxPDa6b PDa6 0 0/4 - 0 0/4 - KU577999 KU577993 FoxPDa9 PDa9 100 3/3 + 0 0/3 - KU577995 KU577989 FoxPDa9b PDa9 0 0/4 - 0 0/4 - KU578000 KU577994 FoxPDa10 PDa10 31 1/4 + 0 0/3 - KU577995 KU577989 FoxPDa13 PDa13 42 1/3 + 0 0/3 - KU577995 KU577989 FoxPDa14 PDa14 56 3/4 + 13 0/4 - KU577995 KU577989 FoxPDa15 PDa15 58 1/3 + 0 0/4 - KU577995 KU577989 FoxPDb1 PDb1 69 2/4 + 0 0/4 - KU577995 KU577989 FoxPDb2 PDb2 69 2/4 + 0 0/4 - KU577995 KU577989 aRDS: Rhizome Discoloration Score bExternal symptoms: number of plants with external symptoms / total number of plants Non-inoculated control plants showed no symptoms.

2.3.2 Assessment of disease suppression

At an inoculum concentration of 5.103 cfu g-1 soil, development of disease symptoms was faster for plants grown in soils from area B than for plants grown in soils from area A. At a concentration of 5.104 cfu g-1 soil, plants grown in soils from area B, as well as the soil from PDa13 and PDa15, rapidly developed disease symptoms. All plants grown in soil from PDa15 showed symptoms at the start of evaluation (Table 2.3).

None of the plants grown in non-inoculated soil showed external symptoms at 94 dpi. Only in one of the five replicates grown in the soil samples from PDa13 and PDb2 minor internal symptoms were observed (RDS 1 and 2, respectively, data not shown).

Disease suppressiveness to Fusarium wilt in an agroforestry system

The level of disease suppression as determined in the greenhouse assay was compiled in a joined rank composed of the ranks of all parameters of the greenhouse assay (namely: the disease index and RAUDPC for both concentrations) (Table 2.3). The patches PDa1, PDa9, PDa10 and PDa14 had a higher degree of disease suppression than the patches PDb1, PDb2, PDa13 and PDa15 (Kruskal Wallis Rank test). The plants on the field were evaluated for production and height. Both factors are summarized in a joined rank (Table 2.3). The results of the greenhouse assay, expressed as the joined rank, correlates positively with the joined rank of the field data (Spearman R2= 0.63; p = 0.0909).

Table 2.3: Determination of suppressive soil patches: the RAUDPC and the disease index of the greenhouse test and the evaluation of the sampled banana cv. Maçã on the field. The joined rank shows the mean of the ranks of the separate variables. Different letters show significant differences between the separate patches (Kruskal Wallis, p = 0.05).

Greenhouse assay Field

5.103 g-1 soil 5.104 g-1 soil

Plant Disease Joined Bunch Joined Soil patch RAUDPCb Disease indexc RAUDPCb height indexc ranka Productiond ranka (m) PDa1 11.2 16.7 (0,0,2) 36.1 58.3 (1,2,4) 6.5 a 4.0 1.0 7.5 PDa9 6.5 16.7 (0,0,2) 24.4 75.0 (2,3,4) 6.9 a 2.0 0.5 4.5 PDa10 13.3 66.7 (1,3,4) 21.7 66.7 (1,3,4) 5.4 a 3.0 1.0 6.5 PDa13 50.0 62.5 (2,3) 57.8 91.7 (3,4,4) 3.3 b 1.0 0.0 1.5 PDa14 11.2 33.3 (0,0,4) 27.1 25.0 (0,1,2) 6.6 a 3.5 1.0 7.0 PDa15 13.3 58.3 (1,2,4) 100.0 100.0 (4,4,4) 3.1 b 2.0 0.5 4.5 PDb1 48.0 100.0 (4,4,4) 74.8 100.0 (4,4,4) 1.9 b 1.2 0.0 2.0 PDb2 38.4 100.0 (4,4,4) 64.0 100.0 (4,4,4) 2.4 b 1.5 0.0 2.5 aThe separate rank assignments can be found in Supplementary material S2.1. The overall rank is the mean of all ranks of the variables. Least suppressive: 1, most suppressive: 8. bRAUDPC: relative area under disease progression curve. The area under disease progress curve of each soil sample can be found in Supplementary material S2.2. cValues between brackets show the Rhizome discoloration score (RDS) of each plant. dBunches with at least 6 hands: 1; bunches with 5 hands or less: 0.5; no bunch production: 0.

Chapter 2

2.3.3 Soil physical-chemical properties

All soil samples were acidic with a pH ranging from 4.6 to 5.6, had a high amount of organic carbon (1.5 to 3.9%) and most had a clay texture, except for patches PDa15 and PDb1, which had a clay loam and sandy clay loam texture, respectively (USDA classification) (Table 2.4).

In the principal component analysis, the first principle component, which explained 61.83% of the total variance, segregated the patches with higher degree of disease suppression from the patches with lower degree of suppression (Figure 2.2). The more suppressive soil patches were characterized by a higher clay content, and a higher pH and EC. The more conducive soils were characterized by a higher sand and silt content, and a higher organic carbon and total nitrogen content. Spearman’s correlation coefficients were significant for pH and texture characteristics (Table 2.4).

Figure 2.2: PCA ordination based on physical-chemical soil characteristics of the soil patches. The four patches with highest level of disease suppression are indicated with green circles and the four most conducive patches with blue triangles. The first principal components explains 61.83% of the variance and the second 23.58%. EC: Electric conductivity, OC: organic carbon, N: total nitrogen.

Disease suppressiveness to Fusarium wilt in an agroforestry system

Table 2.4: Soil physical-chemical characteristics. Spearman’s correlation coefficients with the joined rank for disease suppression as determined in greenhouse assays are shown.

Soil patch OC (%)a TN (%)b pH (KCl) EC (dS m-1)c Sand (%) Silt (%) Clay (%) PDa1 3.34 0.28 5.63 0.14 40.70 16.20 43.10 PDa9 1.58 0.11 5.52 0.11 30.10 11.00 58.90 PDa10 1.54 0.15 5.18 0.08 35.90 15.60 48.50 PDa13 2.01 0.18 5.38 0.12 43.70 14.20 42.20 PDa14 1.62 0.14 5.60 0.12 38.70 15.20 46.10 PDa15 1.89 0.16 5.26 0.09 44.60 20.20 35.20 PDb1 3.91 0.29 4.64 0.05 47.90 19.30 32.80 PDb2 1.97 0.15 4.39 0.06 40.60 17.40 41.90 Correlation with disease -0.57 -0.67 0.81* 0.67 -0.76* -0.76* 0.90* suppression p-value 0.139 0.071 0.015 0.071 0.028 0.028 0.002 aOC: Organic carbon bTN: Total nitrogen cEC: Electric conductivity *: p < 0.05

2.3.4 Vegetation parameters

In a radius of 5 m around the selected banana cv. Maçã the community of neighboring plants, the degree of shadow and the degree of soil cover by graminoids were evaluated (Table 2.5) and analyzed by principal component analysis. The first principal component, which explained 43.33% of the variation, segregated the patches with higher and lower degree of suppression (Figure 2.3). The plant community in patches with a higher degree of suppression had less banana cv. Maçã, a higher diversity of other banana cultivars and a lower soil cover (Table 2.5). The p-value of the spearman correlation between the soil cover and disease suppression was 0.053. Despite being above the 0.05 limit for significance, a strong contribution of the soil cover in the division between suppressive and conducive patches appears from the PCA.

Patches with higher disease suppression did not have fewer banana plants in total, but had a low proportion (max 33%) and lower density (maximum 2 per patch) of cv. Maçã and had at least two other cultivars (Prata, Ouro, Nanica, Terra). Only in the three patches with highest level of disease suppression plants of cv. Ouro were present. In patches with lower degree of disease suppression, cv. Maçã made up at least 75% of the total number of banana plants and at least 3 plants of cv. Maçã per patch were present. Prata, a moderately susceptible cultivar, was the only other cultivar present in those patches (Table 2.5). Figure S2.4 shows the distribution of the banana plants in each patch.

Chapter 2

The three most suppressive patches had a higher abundance of adult trees (at least 2), but had no tree species in common. No adult trees were present in patch PDa10, which had also a high level of disease suppression. The presence of coffee, juvenile trees and shadow did not correlated significantly with the degree of suppression. A table with the plant species per patch can be found in the supplementary data (Table S2.3).

Figure 2.3: PCA ordination based on vegetation parameters per patch: AT: number of adult trees and shrubs, JT: number of juvenile trees and shrubs, C: number of coffee plants, numB: number of banana plants, divB: Shannon diversity of banana cultivars, percM: percentage of banana Maçã, S: shadow, GC: ground cover by graminoids. The four patches with highest level of disease suppression are indicated with green circles and the four most conducive patches with blue triangles. The first principal component explains 43.33% of the variance and the second 20.21%.

Table 2.5: Vegetation characteristics of the patches with different level of disease suppression in soil. Spearman’s correlation coefficients with the joined rank for disease suppression as determined in greenhouse assays are shown.

Banana Number Number cv. cv. cv. cv. Diversity (H’) Soil cover Number adult trees juvenile Soil patch cv. Maçã (HS)a Prata Ouro Terra Nanica in banana Coffee Shadow (by bananas and trees and (MS) (R) (R) (R) cultivars graminoids) shrubs shrubs PDa1 7 2 [29] 0 4 0 1 0.96 4 9 7 2.75 2 PDa9 3 1 [33] 1 1 0 0 1.10 3 2 9 3.25 3 PDa10 5 1 [20] 2 0 0 2 1.05 0 5 24 2.25 2 PDa13 4 3 [75] 1 0 0 0 0.56 1 0 14 4.75 2 PDa14 9 2 [22] 0 5 2 0 1.00 2 5 6 3.75 1 PDa15 5 5 [100] 0 0 0 0 0.00 0 1 1 2.00 4 PDb1 3 3 [100] 0 0 0 0 0.00 1 11 0 3.25 4 PDb2 8 6 [75] 2 0 0 0 0.56 1 7 6 4.00 4 correlation with disease 0.13 -0.79*[-0.72*] -0.03 0.79* 0.41 0.23 0.87* 0.59 -0.29 0.54 -0.06 -0.70 suppression p-value 0.754 0.020 [0.043] 0.952 0.019 0.310 0.577 0.005 0.124 0.490 0.168 0.888 0.053

HS: Highly susceptible to Foc race 1 MS: Moderately susceptible to Foc race 1 R: Resistant to Foc race 1 aThe percentage of cv. Maçã on total number of banana plants per patch is given between square brackets *: p < 0.05

Chapter 2

2.3.5 Soil biotic properties

In the soil samples, culturable bacterial population density varied between 6.08 to 6.65 log cfu g-1 soil and was not correlated with disease suppression. The bacterial community structure, analyzed by DGGE, showed no significant clustering between the soil samples with high or lower level of disease suppression (DGGE profiles in supplementary data S2.5). The profiles were also analyzed for correlation with soil or vegetation properties via redundancy analysis (RDA), but no link was found. The diversity of the most dominant bacterial species, expressed as the number of bands, which mostly represent one species (OTU: operational taxonomic units), was not higher where higher degree of suppression was found. The Fusarium spp. population, analyzed by plating on specific medium, did not differ significantly between the soils with lower and higher level of disease suppression. The patches of area B had a lower density of Fusarium spp. and bacterial diversity than the patches from area A (Table 2.6).

Table 2.6: Microbial characteristics of the patches with different level of disease suppression in soil: culturable populations (CFU) of bacteria on LB medium and Fusarium spp. on Komada medium and the number of operative taxonomic units (OTU) as bands identified with DGGE. Spearman’s correlation coefficients with the joined rank of disease suppression as determined in greenhouse assays are shown.

Culturable Fusarium spp. bacteria Soil patch OTU (log CFU g-1 (log CFU g-1 soil) soil) PDa1 6.49 16 3.16 PDa9 6.38 17 3.33 PDa10 6.13 20 1.62 PDa13 6.47 21 3.30 PDa14 6.08 16 1.02 PDa15 6.65 14 3.22 PDb1 6.20 10 0.60 PDb2 6.30 10 0.70 Correlation with -0.10 0.60 0.57 disease suppression p-value 0.823 0.114 0.139

Disease suppressiveness to Fusarium wilt in an agroforestry system

2.4 Discussion

The objective of this study was to assess levels of disease suppression on a farm in Brazil where Foc disease severity on cv. Maçã was variable and to identify potential soil and plant factors that are correlated with Fusarium wilt suppressiveness. Greenhouse assays with artificial Foc inoculation confirmed that soil samples from various patches in the field differed in levels of disease suppression. Disease suppression was positively correlated with clay, pH, diversity of banana cultivars and cv. Ouro, and negatively correlated with silt, sand, the density of cv. Maçã and soil cover by graminoids.

The conducive area B patches, where banana cv. Maçã was planted in a former pasture, and the most conducive patches of area A had similar plant community properties with a higher density of cv. Maçã and a denser soil cover by graminoids as compared to the more suppressive patches in area A, suggesting a negative impact of these plant factors on disease suppression. Monocotyledons are known to be excellent endophytic carriers of F. oxysporum (Elmer and Lacy, 1986; Gordon et al., 1989; Katan, 1971), and could contribute to inoculum build-up. Also susceptible banana cultivars could be expected to contribute more to inoculum increase than resistant cultivars. However, in our greenhouse assays none of the plants showed external symptoms 3 months after planting in non-inoculated soil, indicating that the natural inoculum level was low in all patches and is negligible compared to artificial inoculation. Thus, in the field, differences in disease are not exclusively due to lower pathogen levels. Susceptible banana cultivars may promote other deleterious rhizosphere microorganisms, which are not directly pathogenic, but affect plants negatively, for instance by their metabolites (Huang et al., 2013; Schippers et al., 1987). From the field observations we cannot determine if the soil cover by graminoids and higher density of cv. Maçã had a negative contribution, or that the more conducive patches just lacked the potential to develop disease suppressiveness.

A mixture of at least two banana cultivars with different levels of Fusarium wilt resistance, beside cv. Maçã, was found on the patches with higher degree of disease suppression. A similar observation was made by Karangwa et al. (2016) who observed that farms in East Africa growing cultivar mixtures had a lower incidence of Fusarium wilt compared to monocultures. Bananas have a wide-spreading root system, and horizontal extension of primary roots can be as far as 5 m, although 1 to 2 m is more common (Robinson and Saúco, 2010). Therefore, there can be considerable overlap and interaction between the root systems of the banana plants considered in the patches, since their distance from the sampled banana plant is maximum 5 m. As banana cultivars differ in their biochemical root

Chapter 2 composition and root exudation, they might be colonized by a different microbial community and modulate the soil microbial community (Buxton, 1962; Wuyts et al., 2007). The interaction of the different cultivars could have led to associational resistance between the neigbouring plants (Saunders et al., 2010). Li et al. (2011a) proposed that plant chemical compounds released in the root environment can contribute to the resistance of banana to Foc. Hopkins et al. (1987) observed that Crimson sweet, a watermelon cultivar moderately resistant to Fusarium wilt, renders the soil suppressive by the stimulation of non-pathogenic F. oxysporum. Similar traits could be present in banana cultivars. The cultivars present on the investigated farm belong to different subgroups and are thus genetically distant. This genetic distance between the cultivars might be a prerequisite to impact the disease when planted in a mixture. Further investigations on cultivar mixtures should be performed to verify the potential beneficial effect on disease suppression and to explore their impact on the pathogen, and the soil and endophytic microbial community. Cultivar Ouro, which was only present in the most suppressive patches, is a promising cultivar for further research.

Stover (1962) already mentioned the potential of mixed plantings to moderate the losses by Fusarium wilt in banana. Remarkably, the variability in disease severity on cv. Maçã was only observed in area A, where a long term agroforestry system was installed. This suggests that growing banana in agroforestry systems may reduce Fusarium wilt disease severity on banana. However, our data do not allow drawing conclusions on this matter, since our study was situated on only one farm and we compared characteristics of patches within the system. At the start of the study, the farmer had proposed a beneficial effect of the tree Trema micrantha on disease suppression, but this could not be confirmed from our data, as we did not find significant correlations between disease suppression and the presence of adult trees and shrubs.

In addition to diversity in banana cultivars, also the higher clay content and pH in the patches with higher disease suppression, may have contributed to the development of disease suppressiveness. Heavy textured soils are generally less prone to Fusarium wilt than sandy soils with a low pH, but some opposite cases have been observed (Amir and Alabouvette, 1993; Baker and Cook, 1974; Domínguez et al., 2001; Höper et al., 1995). Clay soils are considered to have greater disease suppressive potential because they enhance microbial density through their greater ability to retain organic matter (Höper et al., 1995). However, the clay fraction did not correlate positively with organic matter or the density of culturable bacteria. Yet it needs to be mentioned that almost all patches were classified as having a clay texture, so the differences in clay content were small, although significant.

Disease suppressiveness to Fusarium wilt in an agroforestry system

Kinkel et al. (2011) mentioned microbial diversity and density as two important traits to promote the development of disease suppressive soils. However, the patches showing higher disease suppression did not have higher bacterial density, quantified by the total culturable bacteria, nor higher bacterial diversity, as quantified by the DGGE. From the DGGE profile, no common profile of bands was found among the patches with higher disease suppression. A lack of correlation between density and diversity and disease suppression may indicate that specific, rather than general suppressiveness has developed (Höper et al., 1995; Postma et al., 2008). Identification of organisms responsible for soil suppressiveness was beyond the scope of this study.

The presence of Foc on the field can be irregular because of the nature of its dissemination. The farmer could not provide exact information of the origin of all banana cv. Maçã on the field. As it is likely that the infestation occurred through infected planting material and caused a patchy pattern in the spread of the pathogen on the field, we considered only patches with confirmed presence of the pathogen with identical IGS/EF-1α sequences. The pathogen was isolated from the base of one of the pseudostems in the mat, as quantification from soil is not possible because the pathogen cannot be easily distinguished from non-pathogenic F. oxysporum.

To conclude, this study describes a farm, managed as an agroforestry system where different degrees of suppression of Fusarium wilt were found. Improvement of disease suppression in soil is a desirable trait for the management of Fusarium wilt in banana. This study contributes to this objective because potential factors that can improve suppression of Fusarium wilt were identified. Higher diversity of different banana cultivars, presence of cv. Ouro, lower density of the susceptible cv. Maçã and less dense soil cover of graminoids were attributes of sites with higher degree of suppressiveness and can be included in the management and design of a farm. Also a higher clay content was found in suppressive patches, but this trait cannot easily be changed by management. In the field, soil and plant factors occur together and the importance of each individual factor on disease suppression could not be determined from this study. Therefore, more detailed studies are needed to gain further inside in the individual and combined effects of soil and plant factors on Fusarium wilt disease suppression.

Chapter 2

2.5 Acknowledgements

Many thanks are extended to the farmer Geraldo Cândido da Silva and Maria Aparecida de Almeida Pedrosa, whose attentiveness initiated this case study. Their hospitality and help were indispensable to accomplish this work. The authors thank the greenhouse and laboratory technicians, the CTA-ZM and Pieter Cremelie for his assistance with the DGGE. This work is also the fruit of the contributions of Nelleke De Weerd, Diane Chavassieux, Rocío Llamas Vacas and Ellen Velkeneers. The study was funded by a PhD scholarship to P.D. from the Fund for Scientific Research-Flanders (FWO).

2.6 Supplementary material

Table S2.1: Ranks of the indicators of suppressiveness: the RAUDPC and the disease index of the greenhouse test and the evaluation of the sampled banana cv. Maçã on the field.

Greenhouse assay Field 5.103 g-1 soil 5.104 g-1 soil Soil RAUDPC Disease RAUDPC Disease Joined Plant Bunch Joined Patch index index rank height production rank PDa1 6.5 7.5 5 7 6.50 8 7 7.5 PDa9 8 7.5 7 5 6.88 4.5 4.5 4.5 PDa10 4.5 3 8 6 5.38 6 7 6.5 PDa13 1 4 4 4 3.25 1 2 1.5 PDa14 6.5 6 6 8 6.63 7 7 7.0 PDa15 4.5 5 1 2 3.13 4.5 4.5 4.5 PDb1 2 1.5 2 2 1.88 2 2 2.0 PDb2 3 1.5 3 2 2.38 3 2 2.5

Disease suppressiveness to Fusarium wilt in an agroforestry system

3 3 3 PDa1 PDa9 PDa10 2 PDa13 2 2 2 1 1 1 1

0 0 0 0 46 52 61 70 77 84 91 98 46 52 61 70 77 84 91 98 46 52 61 70 77 84 91 98 46 52 61 70 77 84 91 98

3 3 3 PDa14 3 PDa15 PDb1 PDb2 2 2 2 2

1 1 1 1

0 0 0 0 46 52 61 70 77 84 91 98 46 52 61 70 77 84 91 98 46 52 61 70 77 84 91 98 46 52 61 70 77 84 91 98

Figure S2.2a: The progression of the disease incidence in the different soil samples with artificial inoculation with 5*103 g-1 soil. (n = 3). The x-axis shows the days post inoculation and the y-axis the number of diseased plants.

3 3 3 3 PDa1 PDa9 PDa10 PDa13 2 2 2 2

1 1 1 1

0 0 0 0 46 52 61 70 77 84 91 46 52 61 70 77 84 91 46 52 61 70 77 84 91 46 52 61 70 77 84 91

3 3 3 3 PDa14 PDa15 PDb1 PDb2 2 2 2 2

1 1 1 1

0 0 0 0 46 52 61 70 77 84 91 46 52 61 70 77 84 91 46 52 61 70 77 84 91 46 52 61 70 77 84 91

Figure S2.2b: The progression of the disease incidence in the different soil samples with artificial inoculation with 5*104 g-1 soil. (n = 3). The x-axis shows the days post inoculation and the y-axis the number of diseased plants.

Chapter 2

Table S2.3: Perennial plants present in a 5 m radius area around the sampled banana cv. Maçã.

Soil Species (popular Species (Scientific name) Family Number Category Patch name) PDa1 Coffee Coffea sp. Rubiaceae 7 C Banana cv. Ouro Musa sp. Musaceae 4 B Banana cv. Maçã Musa sp. Musaceae 2 B Banana cv. Nanica Musa sp. Musaceae 1 B Crindiúva Trema micrantha Cannabaceae 1 AT Acerola Malpighia glabra Malpighiaceae 1 AT Coqueiro Cocos nucifera Arecaceae 1 AT Leiteira Tabernaemontana salzmannii Apocynaceae 2 1 AT/ 1 JT Marianeira Acnistus arborescens Solanaceae 1 JT Manga Mangifera indica L. Anacardiaceae 1 JT Canela Ocotea sp. Lauraceae 4 JT Jaca Artocarpus heterophyllus Moraceae 1 JT Amexa Prunus sp. Rosaceae 1 JT PDa9 Coffee Coffea sp. Rubiaceae 9 C Banana cv. Ouro Musa sp. Musaceae 1 B Banana cv. Maçã Musa sp. Musaceae 1 B Banana cv. Prata Musa sp. Musaceae 1 B Cabelundinha Myrciaria glazioviana Myrtaceae 1 AT Bico de papagaia Euphorbia pulcherrima Euphorbiaceae 1 AT Abacate Persea americana Lauraceae 1 AT Assa-Peixe Vernonia polysphaera Asteraceae 2 JT PDa10 Coffee Coffea sp. Rubiaceae 24 C Banana cv. Prata Musa sp. Musaceae 2 B Banana cv. Maçã Musa sp. Musaceae 1 B Banana cv. Nanica Musa sp. Musaceae 2 B Canela Ocotea sp. Lauraceae 4 JT Abacate Persea americana Lauraceae 1 JT PDa13 Coffee Coffea sp. Rubiaceae 14 C Banana cv. Maçã Musa sp. Musaceae 3 B Banana cv. Prata Musa sp. Musaceae 1 B Manga Mangifera indica L. Anacardiaceae 1 AT PDa14 Coffee Coffea sp. Rubiaceae 6 C Banana cv. Ouro Musa sp. Musaceae 5 B Banana cv. Maçã Musa sp. Musaceae 2 B Banana cv. Terra Musa sp. Musaceae 2 B Graviola Annona muricata Annonaceae 1 AT Cacau Theobroma cacao Malvaceae 1 AT Goiaba Psidium guajava Myrtaceae 1 JT Mamona Ricinus communis Euphorbiaceae 1 JT Mandioca Manihot esculenta Euphorbiaceae 2 JT Taioba Xanthosoma sagittifolium (L.) Schott, Araceae 1 other PDa15 Coffee Coffea sp. Rubiaceae 1 C Banana cv. Maçã Musa sp. Musaceae 5 B

Disease suppressiveness to Fusarium wilt in an agroforestry system

Assa-peixe Vernonia polysphaera Asteraceae 1 JT PDb1 Banana cv. Maçã Musa sp. Musaceae 3 B Angico vermelho Piptadenia rigida Fabaceae 1 AT Crindiúva Trema micrantha Cannabaceae 7 JT Cabiúna Machaerium incorruptible Fabaceae 4 JT PDb2 Coffee Coffea sp. Rubiaceae 6 C Banana cv. Maçã Musa sp. Musaceae 6 B Banana cv. Prata Musa sp. Musaceae 2 B Angico vermelho Piptadenia rigida Fabaceae 1 AT Assa-peixe Vernonia polysphaera Asteraceae 1 JT Goiaba Psidium guajava Myrtaceae 1 JT Ipê amarela Tabebuia sp. Bignoniaceae 1 JT Pau-Jacaré Piptadenia gonoacantha Fabaceae 1 JT Garapa Apuleia leiocarpa Fabaceae 2 JT Mandioca Manihot esculenta Euphorbiaceae 1 JT Category: C: coffee, B: banana, AT: adult tree or shrub, JT: juvenile tree or shrub.

Chapter 2

Figure S2.4: Position of banana plants in the different patches (circle radius is 5 m)

Figure S2.5: Cluster analysis of the denaturing gradient gel electrophoresis of the 16S rRNA fragments fingerprints of the bacterial community of the different soil samples. The most suppressive soil patches are indicated with circles and most conducive patches with triangles.

Chapter 3

Comparative analysis of pathogenic and non-pathogenic Fusarium oxysporum populations associated with banana on a farm in Minas Gerais

Authors Pauline Deltoura,b, Soraya C. Françaa, Lisa Heymana, Olinto Liparini Pereirab, Monica Höftea

Plant Pathology, in press

Chapter 3

Abstract

Fusarium wilt is one of the most devastating diseases on banana. The causal agent, Fusarium oxysporum f. sp. cubense (Foc) is genetically diverse and its origin and virulence are poorly understood. In this study, pathogenic Foc isolates and non-pathogenic F. oxysporum isolates from Minas Gerais in Brazil were compared using EF-1α and IGS sequences. This allowed us to examine the origin and evolutionary potential of Foc in a country outside the region of origin of the banana plant. Two different sequence types were found among Foc isolates. One sequence type appeared to be of local origin since it was identical to the sequence type of the largest group of non-pathogenic isolates. To explore if the ‘local’ Foc isolates had acquired pathogenicity either independently through coevolution with the host, or through horizontal gene transfer (HGT) of pathogenicity genes from other, probably introduced, Foc isolates, the presence and sequence of putative SIX effector genes were analyzed. Homologues of SIX1, SIX3 and SIX8 were found. SIX1 sequences were identical and exclusively found in all pathogenic isolates, while variable ratios of sequences of multi-copy gene SIX8 were found among non-pathogenic and different pathogenic isolates. This observation supports the HGT hypothesis. Horizontal transfer of genes between isolates of F. oxysporum has important implications on the development of reliable diagnostic tools and effective control measures. Full genome sequencing is required to confirm HGT and to further unravel the virulence mechanisms of forma specialis cubense.

Comparative analysis of Foc and non-pathogenic F. oxysporum populations

3.1 Introduction

Fusarium oxysporum is an ubiquitous fungus found in soils and plants all over the world (Lori et al., 2004). The Fusarium oxysporum species complex (FOSC) is best known for its wilt causing members. Pathogenic F. oxysporum often display a narrow host specificity and are subdivided in formae speciales according to the affected host. However, the species complex also harbors large populations of putative non-pathogenic F. oxysporum, which seem to surpass pathogenic strains in genetic diversity (Lori et al., 2004).

Fusarium oxysporum f. sp. cubense (Foc) is the causal agent of Fusarium wilt on banana. Based on pathogenicity tests, Foc is divided in different races that attack different groups of banana cultivars. Race 1 affects, among others, bananas of the Silk (AAB) and Gros Michel (AAA) cultivars, while race 2 affects cultivar Silver Bluggoe (ABB) and other cooking bananas. Bananas susceptible to race 1 and 2, together with bananas from the Cavendish (AAA) subgroup are affected by race 4. Race 4 is subdivided in a subtropical and tropical race 4, depending on the climatic condition in which they can affect bananas of the subgroup Cavendish. While race 1 is found in almost all banana-growing regions in the world, race 4 is largely restricted to South-East-Asia and Australia, but the race is spreading and has already been reported in a few other Asian countries and Mozambique (Ordonez et al., 2015). Many different clonal lineages of Foc are known, which do not form a monophyletic clade (O’Donnell et al., 1998).

Both race 1 and race 2 are reported in Brazil (Pereira et al., 2005; Ploetz, 1990), but the diversity of Foc seems to differ largely between regions. In Santa Catarina, all Foc isolates collected by Cunha et al. (2015) had an identical elongation factor 1α sequence type (EF-1α). Another study, comprising isolates from several states in Brazil showed a higher diversity, but with dominance of two microsatellite profiles, indicating a clonal structure of the pathogen (Costa et al., 2015).

The genome of F. oxysporum can be divided in a core and a lineage specific region. The core genomic regions reside on conserved chromosomes and are present in all F. oxysporum strains, regardless of their hosts (Ma, 2014). The lineage specific genomic regions are variable between strains, contain no housekeeping genes and are generally rich in transposons. In the tomato pathogen, Fusarium oxysporum f. sp. lycopersici strain Fol4287, the lineage specific region consists out of four chromosomes and two smaller regions at the end of core chromosomes. One of the lineage specific chromosomes contains a high fraction of effector genes, and is largely responsible for its pathogenicity, as horizontal

Chapter 3 transfer of this pathogenicity chromosome to a non-pathogenic strain, Fo47, turned it virulent to tomato (Ma et al., 2010). One major group of effector genes on this pathogenicity chromosome is formed by the SIX (secreted in xylem) genes. They code for small, commonly cysteine-rich proteins that lack homology to other proteins, contain a signal peptide for secretion and are probably involved in the interaction with the plant during infection (de Sain and Rep, 2015). To date, 14 SIX genes have been described in F. oxysporum f. sp. lycopersici (Lievens et al., 2009, Schmith et al., 2013), four of which are needed for full virulence on tomato, namely SIX1, SIX3, SIX5 and SIX6 (Gawehns et al., 2015). Some SIX genes are involved in gene-for-gene interactions with the tomato plant (de Sain and Rep, 2015). Homologues of SIX genes have been detected in F. oxysporum pathogenic on several hosts such as cabbage, cotton, onion, Arabidopsis and banana (Chakrabarti et al., 2011; Fraser-Smith et al., 2014; Kashiwa et al., 2013; Meldrum et al., 2012; Sasaki et al., 2015; Thatcher et al., 2013), but their role in virulence on other hosts has not yet been revealed. In Foc, homologues of SIX1, SIX2, SIX4, SIX6, SIX7, SIX8, SIX9, and SIX13 have been reported (Fraser-Smith et al., 2014). The SIX genes have been used to predict host specificity of F. oxysporum (van Dam et al., 2016). Primers have been developed based on SIX1, SIX2 and SIX3 to distinguish F. oxysporum f. sp. lycopersici from other formae speciales. In addition, different races of F. oxysporum f. sp. lycopersici can be identified based on SIX3 and SIX4 (Lievens et al., 2009). Fraser-Smith et al. (2014) developed primers based on SIX8 to discriminate the different races of Foc.

F. oxysporum is a cosmopolitan soil inhabitant, suggesting that it is also indigenous to South- American soils. On the contrary, banana, which natural distribution covers South-East Asia, was probably only introduced in the Americas in the post-Columbian period (De Langhe et al., 2009). Being an introduced crop, the coevolution of banana with indigenous F. oxysporum has probably been too short to result in pathogenic isolates, and it could be suggested that all Foc are introduced in Brazil. By studying the relation of the local non- pathogenic population and Foc, outside the region of origin of the host plant, we aimed to answer questions about the origin and evolutionary potential of Foc.

We characterized a population of F. oxysporum isolates from a farm in Pedra Dourada, Minas Gerais, Brazil. Isolates were collected from soil, graminoids, banana cultivars (cv.) Maçã, Prata and Ouro to explore the occurrence of Foc inside and outside its host. In this study, this collection of isolates which comprised both Foc and strains non-pathogenic to banana was molecularly characterized to study the local population structure and resemblance between pathogenic and non-pathogenic strains. Two marker genes of the core genome, EF-1α and IGS, were used since those have been shown to contain the highest

Comparative analysis of Foc and non-pathogenic F. oxysporum populations nucleotide variation of marker genes used in the F. oxysporum species complex (O’Donnell et al., 2009). The population comprised Foc with both divergent and identical sequence types (St) as the non-pathogenic isolates, suggesting the occurrence of a local pathogenic population. To test if the ‘local’ Foc isolates attained pathogenicity through coevolution with banana or via horizontal gene transfer (HGT) between an introduced pathogen and a native non-pathogenic strain, the presence and sequence of putative SIX effectors was analyzed. Throughout the text, the terms pathogenic and non-pathogenic always refer to the virulence on banana cv. Silk, unless clearly mentioned.

3.2 Material and methods

3.2.1 F. oxysporum isolates

Samples were collected at a farm in Pedra Dourada, MG, Brazil (20°50’ S, 42°07’ W). The farm is managed as an agroforestry system with banana and coffee as main crops, which are intercropped with various fruit and native tree species. On the farm, banana cv. Maçã (Subgroup: Silk) was affected by Fusarium wilt, but disease was variable and variation could be attributed to differences in disease suppressiveness of the soil (Chapter 2). Soil samples, different graminoids, pseudostem and roots of banana cultivars (cv.) Ouro (AA, Sucrier subgroup, resistant to Foc race 1) and Prata (AAB, Pome subgroup, moderately susceptible to Foc race 1) and pseudostem of banana Maçã (AAB, Silk subgroup, very susceptible to Foc race 1) were collected on the farm for isolation of F. oxysporum (Table 1). Isolates were also recovered from roots, rhizome and pseudostem of tissue culture-derived plantlets of banana cv. Maçã (Multiplanta Tecnologia Vegetal) grown in soil sampled from the farm. In addition, diseased banana cv. Maçã from two farms in Araponga (approximately 42 km from Pedra Dourada, straight distance) were sampled. All plants sampled were evaluated for internal discoloration.

For fungal isolation, plant samples were surface sterilized in NaOCl, washed in sterile water and plated on Potato Dextrose Agar (PDA), supplemented with 10 mg l-1 rifampicin (for 1 l PDA: an infusion of 200 g potatoes, 20 g dextrose, 15 g agar, complemented with distilled water). Soil samples were serially diluted in sterile water and plated on rose Bengal agar. Isolates morphologically identified as F. oxysporum were kept as single spore culture on dried filter paper and slants with PDA.

Chapter 3

3.2.2 Pathogenicity test on banana

The isolates collected from the pseudostem of cv. Maçã in Pedra Dourada were tested for pathogenicity at the Federal University of Viçosa. Tissue culture-derived plantlets of cv. Maçã (AAB, subgroup Silk, susceptible to Foc race 1) and cv. Grande Naine (AAA, subgroup Cavendish, susceptible to Foc race 4) (Multiplanta Tecnologia Vegetal) were dipped during 2 h in a spore suspension (106 spores ml-1) prepared from 10-day-old cultures on PDA and grown for 7 weeks in the greenhouse (18-28°C average night-day temperature; 12:12 light regime), before evaluating external (yellowing of the leaves, splitting of the pseudostem and stunted growth) and internal symptoms. The vascular discoloration of the rhizome was given a score (rhizome discoloration score: RDS 0: 0%; 1:1-25%; 2: 26-50%; 3: 51-75%; 4: 76- 100%). When a score different from 0 was given, the isolate was considered pathogenic.

All other isolates were tested for pathogenicity at the lab of Phytopathology, Ghent University, Belgium. Eight-week-old tissue culture-derived plantlets of cv. Silk (accession number ITC0348) were dipped for 2 h in a spore suspension (5. 105 spores ml-1), obtained from a 2- week-old culture grown on PDA. Plantlets were grown in potting soil (Universal, type 1, M. Snebbout n.v./s.a.) in a growth room (26°C; 16:8 day/night light regime). After 5 weeks, plantlets were uprooted and internal symptoms were evaluated as described above.

3.2.3 Pathogenicity test on tomato

Pathogenicity on tomato was tested for a subset of the Foc and non-pathogenic isolates following the protocol of van der Does et al. (2008a). Four F. oxysporum f. sp. lycopersici 168 (Fol) isolates (race 1, 2 and 3) were included as control. Ten-day-old plantlets of cultivar Moneymaker were uprooted and dipped in a spore suspension of 107 spores ml-1 for 30 min, made from a 2-week-old culture on PDA. Plantlets were planted in 200 ml pots with potting soil (Universal, type 1, M. Snebbout n.v./s.a.) and grown in a growth chamber (25°C, 16:8 day/night light regime) for 3 weeks until evaluation. A disease score from 0 to 4 was attributed depending on the extent of browning of vessels: 0, no symptoms; 1, slightly swollen and or bent hypocotyl; 2, one or two brown vascular bundles in hypocotyl; 3, at least two brown vascular bundles and growth distortion; 4, all vascular bundles are brown, plant either dead or very wilted.

Comparative analysis of Foc and non-pathogenic F. oxysporum populations

3.2.4 DNA extraction, PCR and sequencing

Isolates were grown for 10 days on potato dextrose broth (BD Difco), crushed in liquid nitrogen and DNA extracted by use of the Wizard Genomic DNA purification kit (Promega), following the protocol of the manufacturer. DNA samples were kept at -20°C until use. The translation elongation factor- 1α (EF-1α) gene was amplified using the primer EF1 and EF2 and the program described by O’Donnell et al. (1998). The intergenic spacer region (IGS) was amplified using primers iNL1 and iCNS1 and following program: 95°C for 5 min and 30 cycles of 95°C for 1 min, 62°C for 1 min and 72°C for 3 min, followed by an additional extension time for 10 min at 72°C (O’Donnell et al., 2009). After purification with Exosap-IT® (Affymetrix), PCR products were sent for sequencing to LGC genomics, Berlin, Germany with the primers mentioned above. For the sequencing of the IGS sequence, the samples were sent with two additional primers: NLa and CNSa to obtain reliable coverage for the entire 2200 bp (O’Donnell et al., 2009).

A subset of the isolates was screened for presence of SIX homologues. Primer pairs and amplification conditions for SIX1 to SIX8 were used as described by Meldrum et al. (2012). The primers and amplification conditions for SIX9 to SIX11 and SIX14 were used from Laurence et al. (2015). Only partial fragments of the SIX genes were obtained. When a clear band of the expected length was obtained, the PCR product was purified with Exosap-IT® (Affymetrix), following the manufacturers guidelines, and sent for sequencing (LGC genomics).

3.2.5 Data analysis

The isolates were confirmed to belong to the F. oxysporum species complex by comparing the EF-1α sequences to reference sequences from Genbank and Fusarium-ID. Additionally, IGS and EF-1α sequences were compared to the multi locus sequence types (MLST) as described in O’Donnell et al. (2009) and listed in Fusarium-ID to identify new MLST. The quality of all sequences were manually checked. Sequence types, which were determined including gaps as informative characters, were attributed a number (EF-1α) and letter (IGS) (Table 3.1). Sequences identified in this work have been deposited to GenBank (Table 3.1).

Sequences were aligned by use of the CrustalW multiple alignment in BioEdit, and manually edited afterwards. Maximum likelihood analysis was performed in MEGA6 on all sequence types, including sequences of the MLST described by O’Donnell et al. (2009) that contain

Chapter 3 isolates pathogenic on banana as reference strains. The general time reversible model was applied, with the Nearest-Neighbor-Interchange as heuristic method and partial deletion as option for gaps. Support for each branch was assessed with 1000 bootstrap replicates. Two isolates of the sister taxon of the F. oxysporum species complex, F. foetens were used to root the tree (O’Donnell et al., 2009).

Sequences of putative SIX genes were aligned against SIX homologues found in different formae speciales available on GenBank and from the Broad institute of MIT and Harvard (Fusarium comparative genome project).

3.2.6 Carbon-utilization pattern

The growth profile on several sugars and amino acids was characterized for a subset of the isolates based on the method derived from Steinberg et al. (1999), initially to test if pathogenic isolates shared a common growth profile on proteins. In a 96-well plate, 200 µl of −1 minimal medium (l distilled water : Na2NO3 2 g, KH2PO4 1 g, MgSO4.7H2O 2.5 g, KCl 0.5 g, trace element solution 2 ml. Trace element solution composition was, l−1 distilled water: citric acid 5 g, ZnSO4.7H2O 5 g, FeSO4.7H2O 4.75 g, Fe(NH4)2(SO4)2.6H2O 1 g, CuSO4.5H2O 0.25 -1 g, MnSO4.H2O 0.05 g, H3BO3 0.05 g, NaMoO4.2H2O 0.05 g) was supplemented with 5 g l of a specific carbon source and inoculated with 50 µl of a spore suspension (2. 105 spore ml-1) obtained from 14-day-old cultures on PDA. Per carbon source, three wells were supplemented with the same isolate. The plates were incubated in the dark at 25°C. After 4 days, the growth was assessed by turbidity measurements at 595 nm using a multiscan EX spectrophotometer (Thermo Labsystems). Hierarchical clustering (Ward’s method, R package pvclust) was performed on the mean optical density per carbon source to study similarities between growth profiles and the approximately unbiased p-values (AU-p value) were calculated.

Comparative analysis of Foc and non-pathogenic F. oxysporum populations

3.3 Results

3.3.1 Isolation of F. oxysporum and pathogenicity on banana

In total, 71 isolates of F. oxysporum were collected (Table 3.1). From the farm in Pedra Dourada, 43 isolates were collected from soil, different graminoids, pseudostem and roots of banana cultivars (cv.) Ouro and Prata and pseudostem of banana cv. Maçã. Symptoms of Fusarium wilt were only observed in cv. Maçã. In addition, 26 isolates were collected from the root, rhizome and pseudostem of tissue culture-derived plantlets of banana cv. Maçã grown for 3 months in the greenhouse in soil collected on different spots on the farm in Pedra Dourada. Finally, two isolates were obtained from the rhizome of clearly affected banana cv. Maçã on two farms in Araponga.

All isolates were tested for pathogenicity on banana cv. Silk, a cultivar susceptible to Foc race 1. Pathogenicity was exclusively observed in isolates obtained from Banana cv. Maçã. None of the isolates collected from a non-host was pathogenic on banana, including the isolates collected from a moderately susceptible (cv. Prata) or resistant (cv. Ouro) banana cultivar. From the farm in Pedra Dourada, nine pathogenic isolates were obtained, which were also tested for pathogenicity on cv. Grande Naine. One of the pathogenic isolates, FoxPDa3, also caused clear symptoms on cv. Grande Naine in the greenhouse assay (see Chapter 2). In addition, four isolates that were recovered from banana cv. Maçã plantlets grown in soil samples from the farm in Pedra Dourada were pathogenic to banana cv. Silk. Also the two isolates obtained from affected cv. Maçã on farms in Araponga were pathogenic to banana cv. Silk, resulting in a total of 15 pathogenic isolates and 56 non-pathogenic isolates (Table 3.1).

3.3.2 Population diversity based on EF-1α and IGS

Sixteen different EF-1α sequence types, named St1 to St16 (approximately 653bp), including insertions/deletions (indels) as informative characters were identified. Forty-eight sites were polymorphic, with 42 single nucleotide polymorphisms and six indels. Of the whole population, 52% belonged to the lineage containing St1 to St5. The sequence types within this lineage differ by maximum one single nucleotide or indel (Figure 3.1).

Chapter 3

Twenty-five different IGS haplotypes, named StA to StY, were found: of the approximately 1922 bp IGS fragment, 285 single nucleotide polymorphisms and 22 indels were found. Topological incongruence was found between the trees based on the EF-1α and IGS sequence. The dominant EF-1α lineage (St1 to St5) segregates in two different monophyletic lineages and two separate isolates in the phylogenetic tree based on the IGS sequences. Also isolates of the lineages St12-St15 and St7, both monophyletic, were segregated in the tree based on IGS (Figure 3.1).

We found 32 unique EF-1α/IGS multilocus sequence types (MLST), 22 of which only comprise one isolate. Only EF-1α/IGS sequence types St1/StB, St9/StQ, St7/StS and St16/StP belong to a MLST already described by O’Donnell et al. (2009), respectively MLST 28, 25, 20 and 16.

The pathogenic isolates are divided in two different sequence types. The first group, belonging to St9/StQ, contains all pathogenic isolates from Pedra Dourada with the exception of one and contains no non-pathogenic isolates. The second group, belonging to St1/StA, contains a single pathogenic isolate from Pedra Dourada (FoxPDa3), and the two isolates from Araponga (Ara1 and Ara2). St1/StA contains many non-pathogenic isolates that form locally the dominant group.

Table 3.1: Strains of the F. oxysporum complex used in this study

Sequence Pathogenicityb Accession numbers types Name Host Plant part Locationa EF-1α IGS Banana Banana Tomato EF-1α IGS cv. Silk cv. cv. Grande Moneymaker Naine Fusarium oxysporum f. sp. cubense

FoxPDa1 Banana cv. Maça Pseudostem PD 9 Q + - - KU577989 KY296389 FoxPDa9 Banana cv. Maça Pseudostem PD 9 Q + - nt KU577989 KY296389 FoxPDa10 Banana cv. Maça Pseudostem PD 9 Q + - nt KU577989 KY296389 FoxPDa13 Banana cv. Maça Pseudostem PD 9 Q + - nt KU577989 KY296389 FoxPDa14 Banana cv. Maça Pseudostem PD 9 Q + - nt KU577989 KY296389 FoxPDa15 Banana cv. Maça Pseudostem PD 9 Q + - - KU577989 KY296389 FoxPDb1 Banana cv. Maça Pseudostem PD 9 Q + - - KU577989 KY296389 FoxPDb2 Banana cv. Maça Pseudostem PD 9 Q + - nt KU577989 KY296389 MrPDa9a Banana cv. Maça Root ePD 9 Q + nt nt KU577989 KY296389 MsPDa9a Banana cv. Maça Rhizome ePD 9 Q + nt nt KU577989 KY296389 MsPDa13 Banana cv. Maça Rhizome ePD 9 Q + nt nt KU577989 KY296389 MsPDb2b Banana cv. Maça Rhizome ePD 9 Q + nt nt KU577989 KY296389 FoxPDa3 Banana cv. Maça Pseudostem PD 1 A + + - KU577991 KY296373 Ara1 Banana cv. Maça Pseudostem Araponga 1 A + nt - KU577991 KY296373 Ara2 Banana cv. Maça Pseudostem Araponga 1 A + nt - KU577991 KY296373 F. oxysporum non-pathogenic on banana cv. Silk

SPDa1 Soil n/a PD 1 A - - nt KU577991 KY296373 SPDa5 Soil n/a PD 11 V - - nt KY274447 KY296394 SPDa7 Soil n/a PD 13 R - - nt KY274449 KY296390 SPDa14a Soil n/a PD 1 B - - nt KU577991 KY296374 SPDa14b Soil n/a PD 3 H - - nt KY274441 KY296380 SPDa9 Soil n/a PD 4 G - - nt KY274442 KY296379 FoxPDa2 Banana cv. Maça Pseudostem PD 1 A - - - KU577991 KY296373

FoxPDa6a Banana cv. Maça Pseudostem PD 10 N - - nt KU577992 KY296386 FoxPDa6b Banana cv. Maça Pseudostem PD 1 A - - - KU577991 KY296373 FoxPDa9b Banana cv. Maça Pseudostem PD 1 C - - - KU577991 KY296375 FoxPDa9c Banana cv. Maça Pseudostem PD 1 I - nt nt KU577991 KY296381 PpPDa9 Banana cv. Prata Pseudostem PD 1 A - nt nt KU577991 KY296373 PrPDa9 Banana cv. Prata Root PD 1 F - nt nt KU577991 KY296378 PrPDa10 Banana cv. Prata Root PD 1 C - nt - KU577991 KY296375 PrPDa11 Banana cv. Prata Root PD 1 B - nt - KU577991 KY296374 PpPDa13 Banana cv. Prata Pseudostem PD 1 C - nt nt KU577991 KY296375 OrPDa14 Banana cv. Ouro Root PD 3 H - nt nt KY274441 KY296380 OrPDa16 Banana cv. Ouro Root PD 2 A - nt nt KY274440 KY296373 OpPDa16 Banana cv. Ouro Pseudostem PD 1 J - nt nt KU577991 KY296382 OrPDa17 Banana cv. Ouro Root PD 5 A - nt nt KY274443 KY296373 GrPDb1a Melinis minutiflora Root PD 7 E - - nt KY274445 KY296377 GrPDb1b Hyparhernia sufa Root PD 14 K - nt nt KY274450 KY296383 GsPDb1c Digitaria insularis Shoot PD 12 O - - nt KY274448 KY296387 GrPDa1a Digitaria insularis Root PD 1 C - - nt KU577991 KY296375 GrPDa1b Cyperus iria Root PD 1 B - - nt KU577991 KY296374 GrPDb2 Brachiaria brizantha Root PD 6 B - - nt KY274444 KY296374 GrPDa12a Hypolytrum Root PD 7 S - nt nt KY274445 KY296391 schradarianum GsPDa12b Digitaria (horizontalis) Shoot PD 1 B - nt nt KU577991 KY296374 GrPDa12b Digitaria (horizontalis) Root PD 1 A - nt nt KU577991 KY296373 GrPDa5 Rhynchospora aurea Root PD 1 A - nt nt KU577991 KY296373 GrPDa15 Paspalum sp. Root PD 2 A - nt nt KY274440 KY296373 GrPDa17a Panicum repens Root PD 2 F - nt nt KY274440 KY296378 GsPDa17b Cyperus sp. Shoot PD 6 B - nt nt KY274444 KY296374 GrPDa17b Cyperus sp. Root PD 13 X - nt nt KY274449 KY296396 MrPDa1a Banana cv. Maça Root ePD 3 H - nt nt KY274441 KY296380 MrPDa1b Banana cv. Maça Root ePD 1 A - nt nt KU577991 KY296373

MrPDa5 Banana cv. Maça Root ePD 1 H - nt nt KU577991 KY296380 MrPDa6 Banana cv. Maça Root ePD 14 K - nt nt KY274450 KY296383 MrPDa7 Banana cv. Maça Root ePD 7 T - nt nt KY274445 KY296392 MrPDa13 Banana cv. Maça Root ePD 1 A - nt nt KU577991 KY296373 MrPDa14a Banana cv. Maça Root ePD 8 D - nt nt KY274446 KY296376 MrPDa14b Banana cv. Maça Root ePD 14 M - nt nt KY274450 KY296385 MrPDa15 Banana cv. Maça Root ePD 1 C - nt nt KU577991 KY296375 MrPDb1 Banana cv. Maça Root ePD 11 W - nt nt KY274447 KY296395 MrPDb2 Banana cv. Maça Root ePD 7 E - nt nt KY274445 KY296377 MpPDa1 Banana cv. Maça Pseudostem ePD 1 G - nt nt KU577991 KY296379 MpPDa10 Banana cv. Maça Pseudostem ePD 1 B - nt nt KU577991 KY296374 MsPDa4 Banana cv. Maça Rhizome ePD 1 C - nt nt KU577991 KY296375 MsPDa7 Banana cv. Maça Rhizome ePD 1 A - nt nt KU577991 KY296373 MsPDa9b Banana cv. Maça Rhizome ePD 1 F - nt nt KU577991 KY296378 MsPDa10 Banana cv. Maça Rhizome ePD 15 L - nt nt KY274451 KY296384 MsPDa14b Banana cv. Maça Rhizome ePD 2 A - nt nt KY274440 KY296373 MsPDa15 Banana cv. Maça Rhizome ePD 2 C - nt nt KY274440 KY296375 MsPDb1a Banana cv. Maça Rhizome ePD 8 Y - nt nt KY274446 KY296397 MsPDb1b Banana cv. Maça Rhizome ePD 7 U - nt nt KY274445 KY296393 MsPDb2a Banana cv. Maça Rhizome ePD 16 P - nt nt KY274452 KY296388 Fusarium oxysporum f. sp. lycopersici

FolR1 Tomato Brazil nt nt + (race 1)

BFOL-70 Tomato Louisiana, USA nt nt + (race 2) IPO3 Tomato The Netherlands nt nt + (race 3) FOL- Tomato Arkansas, USA nt nt + (race 3) MM10 a PD: Pedra Dourada, ePD: experiment with soil samples from Pedra Dourada b nt: not tested

Chapter 3

Figure 3.1: Trees with the highest log likelihood based on the EF-1α and IGS sequence that were generated using the Maximum Likelihood method, using 1000 bootstraps. The tree is rooted with F. foetens isolates NRRL 38302 and NRRL 31852. The numbers at the nodes show the percentage of trees in which the associated taxa clustered together. The Multi locus sequence types (MLST) containing f. sp. cubense as described by O’Donnell et al. (2009) were added as reference strains. Strains are grouped in their respective sequence types and strains pathogenic to banana cv. Silk are indicated in red and underlined. Colored boxes show the grouping of the isolates based on the EF-1α allele and their respective position on the tree based on the IGS allele. Well supported lineages (bootstrap > 70) are indicated in bold.

Comparative analysis of Foc and non-pathogenic F. oxysporum populations

3.3.3 SIX genes

A subset of the population comprising six Foc isolates (three isolates of St9/StQ and three from St1/StA) and six non-pathogenic F. oxsporum isolates (three isolates of St1/StA, one of St1/StB and two of St1/StC) was screened for homologues of the SIX genes. Isolates of F. oxysporum f. sp. lycopersici (Fol) race 2 and 3 were included as reference, and resulted in clear bands of the expected length for all SIX genes, except SIX4. Only of SIX1, SIX3 and SIX8 clear amplicons with identical length as the references were observed (Table 3.2). Other primers resulted in multiple non-specific PCR products (SIX2, SIX5 , SIX6 , SIX7, SIX9, SIX10, SIX11 and SIX14) and no amplicon was found after amplification with SIX4 primers.

Table 3.2: Screening for SIX genes in a subset of the isolates

PCR amplicon Accession numbers Isolate SIX1 SIX3 SIX8 SIX1 SIX3 SIX8 Fusarium oxysporum f. sp. lycopersici BFOL-70 + + + IPO3 + + + Fusarium oxysporum f. sp. cubense FoxPDa15 + + + KY296398 KY296399 KY296401 FoxPDb1 + + + KY296398 KY296399 KY296401 FoxPDa1 + + + KY296398 KY296399 KY296401 Ara1 + + + KY296398 KY296400 KY296401 Ara2 + + + KY296398 KY296399 KY296401 FoxPDa3 + - + KY296398 KY296403 Non-pathogenic Fusarium oxysporum FoxPDa2 - - + KY296402 FoxPDa6b - + - KY296400 PpPDa9 - - - FoxPDa9b - - + KY296401 PrPDa10 - - - PrPDa11 - + - KY296400

PCR amplification with the SIX1 primers resulted in a 260 bp band for the Fol controls and all Foc isolates. No band with similar length was found in the non-pathogenic isolates. The nucleotide sequences of SIX1 of all Foc isolates were identical (Accession number: KY296398), and were identical to SIX1 of Australian Foc isolates described by Meldrum et al. (2012) and Laurence et al. (2015).

Chapter 3

Amplification with the SIX3 primers showed a 555bp fragment for the Fol controls and for all Foc isolates, except for FoxPDa3. An amplicon was also observed for two non-pathogenic isolates (FoxPDa6b and PrPDa11). Two different homologues were found. A first homologue (Accession number: KY296399) was found in the strains FoxPDa1, FoxPDa15, FoxPDb1 and Ara2 and was identical over the full length to the sequences of F. oxysporum f. sp. lycopersici (Fol) race 3 isolate IPO3. The second homologue (Accession number: KY296400), found in isolates Ara1, FoxPDA6b and PrPDa11 was identical to the sequences of Fol race 3 isolate FOL-MM10 (Figure 3.2). To our knowledge, to date these SIX3 sequences have been exclusively found in Fol isolates, and have been used for race determination within Fol. Despite the presence of homologues of SIX3 identical to sequences found in Fol race 3, the isolates caused no vascular discoloration or growth distortions in tomato plants cv. Moneymaker, which is susceptible to all Fol races (Figure 3.3).

BFOL-51 -GACGGGGTAACCCATATTGCGTGTTTCCCGGCCGCCGCACGTCTTCTACT- 14844 -GACGGGGTAACCCATATTGCGTGTTTCCCGGCCACCGCACGTCTTCTACT- IPO3 -GACGGGGTAACCCATATTGCGTGTTTCCCGGCCGCCCCACGTCTTCTACT- FOL-MM10 -GACGGGGTAACCCATATTGCATGTTTCCCGGCCGCCGCACGTCTTCTACT- FoxPDb1a -GACGGGGTAACCCATATTGCGTGTTTCCCGGCCGCCCCACGTCTTCTACT- FoxPDa6bb -GACGGGGTAACCCATATTGCATGTTTCCCGGCCGCCGCACGTCTTCTACT-

Figure 3.2: Sequence alignments of part of the SIX3 genes where SNP’s were found. Comparison of isolates from this study with F. oxysporum f. sp. lycopersici (Fol) strains from Lievens et al. (2009). BFOL-51 shows the sequence of Fol race 1 and race 2, while isolates 14844, IPO3 and FOL-MM10 represent the different sequence types of Fol race 3. Variable nucleotides are indicated in bold and marked in grey. a Isolates FoxPDa15, FoxPDa1 and Ara2 have the same sequence as FoxPDb1. b Isolates Ara1 and PrPDa11 have the same sequences as FoxPDa6b.

Comparative analysis of Foc and non-pathogenic F. oxysporum populations

7 6 5 4 4 3 2 3 1

Numberofplants 2 0

Ara1 Ara2

IPO3 FolR1

Water 0

BFOL-70

PrPDa11 PrPDa10

FoxPDa1 FoxPDa3 FoxPDa2

FoxPDb1

FoxPDa15

FoxPDa6b FoxPDa9b FOL-MM10 Ctrl - Ctrl + SIX3 + SIX3 -

Figure 3.3: Disease scores of the different strains on tomato plants, cv. Moneymaker (susceptible to all races of F. oxysporum f. sp. lycopersici). Isolates FolR1, BFOL-70, IPO3 and FOL-MM10, respectively F. oxysporum f. sp. lycopersici race 1, 2, 3 and 3, were included as controls.

A homologue of SIX8 was detected in all tested pathogenic strains and in two non- pathogenic strains (FoxPDa2 and FoxPDa9b). The 233 bp amplicon stretches along the 3rd exon as proposed by Fraser-Smith et al. (2014). At least two homologues of multi-copy gene SIX8 were found. The pathogenic strains FoxPDb1, FoxPDa1, FoxPDa15, Ara1, Ara2 and non-pathogenic strain FoxPDa9b shared the same unique sequence (Accession number: KY296401, Figure 3.4). This sequence showed high similarity with a SIX8 sequence of Fol isolate 4287 (97.9% homology; Figure 3.5). Two different sequence types seemed to be present in non-pathogenic strain FoxPDa2 (Accession number: KY296402), but in different ratio: a high copy number of a sequence highly similar to homologues of SIX8 observed in Foc subtropical race 4 (Meldrum et al., 2012) and the homologues described as Foc-SIX8a in Fraser-Smith et al. (2014), and in a low copy-number, the sequence type found in FoxPDb1 (Figure 3.4). In pathogenic isolate FoxPDa3 a high frequency of ambiguous bases was found (Figure 3.4 and Figure 3.5). The sequence of FoxPDa3 (Accession number: KY296403) appears to correspond to an about equal copy number of the main sequence of FoxPDa2 and the sequence of the other isolates (Figure 3.4).

Chapter 3

Figure 3.4: Illustration of the electropherogram showing ambiguous base pairs in a SIX8 fragment, bp 149-218, of the isolates FoxPDb1 (Foc, group St9/StQ), FoxPDa2 (non-pathogenic F. oxysporum), FoxPDA3 (Foc, group St1/StA). Blue arrows indicate clear ambiguous base pairs, grey arrows show corresponding base pairs with little or no ambiguity.

Comparative analysis of Foc and non-pathogenic F. oxysporum populations

1 - 61

W95-182 G C C T G C A T A A C A G G T G C C G G C G T G G C G T T T C A G G C A T A T G C T G G A T G C T A T C T C A C T G C A T INDO 155 G C C T G C A T A A C A G G T G C T G G C G T G G C G T T T C A G G C A T A T G C T G G A T G C T A T C T C A C T G C A T 22615 G C C T G C A T A A C A G G T G C C G G T G T G G C G T T T C A G G C A T A T G C T G G A T G C T A T C T C A C T G C A T Fol4287 T A G T A T A T A A C A G G T G C C G G T T C G G C G T T T C A G G C A T A T G C A G G A T G C T A T C T C A C T G C A T FoxPDa2 G C C T G C A T A A C A G G T G C C G G Y K T G G C G T T T C A G G C A T A T G C W G G A T G C T A T C T C A C T G C A T FoxPDb1 G C C T G C A T A A C A G G T G C C G G T T C G G C G T T T C A G G C A T A T G C A G G A T G C T A T C T C A C T G C A T FoxPDa3 G C C T G C A T A A C A G G T G C C G G Y K Y G G C G T T T C A G G C A T A T G C W G G A T G C T A T C T C A C T G C A T

62 - 122

W95-182 T T C G T A A T G A C C C G C G C A C T C T T A C G T T G A G G A T G G A C A G A A C T A G G G G C G G A G G G A C A T C INDO 155 T T C G T A A T G A C C C G C G C A C T C T T A C G T T G A G G A T G G A C A G A A C T A G G G G C G G A G G G A C A T C 22615 T T C G T A A Y G A C C C G C G C A C T C T T A C G T T G A G G A T R G A C A G A A C T A G G G G C G G A G G G A T A T C Fol4287 T T C G T A A T G A C C C G C G C A C T C T T A C G T T G A G G A T G G A C A A A A C T A G G G G T G A A C G G A T A T C FoxPDa2 T T C G T A A T G A C C C G C G C A C T C T T A C G T T G A G G A T G G A C A G A A C T A G G G G C G G A G G G A C A T C FoxPDb1 T T C G T A A T G A C C C G C G C A C T C T T A C G T T G A G G A T G G A C A A A A C T A G G G G T G A A C G G A T A T C FoxPDa3 T T C G T A A T G A C C C G C G C A C T C T T A C G T T G A G G A T G G A C A R A A C T A G G G G Y G R A S G G A Y A T C

123 - 184

W95-182 G A A C G T G C T T G T C C T T T T G A G C G G G G G A G C G C T A T C A C A T G C G G T A C G G G A A G T A G T C C A G A INDO 155 G A A C G T G C T T G T C C T T T T G A G C G G G G G A G C G C T A T C A C A T G C G G T A C G G G A A G T A G T C C A G A 22615 G A A C G T G C T T G T T A T T T T G A G C G G G G G A G C G C T M T C A C A T G C G G T A G A G G A G G T A G T C C G A A Fol4287 G A A C G T G C T T G T T A T T T T G A G C G G G G G A G C G C T A T C A C A T G C G G T G G A G G A G G T A G T C C A A A FoxPDa2 G A A C G T G C T T G T C C T T T T G A G C G G G G G A G C G C T M T C A C A T G C G G T A C R G G A A G T A G T C C A G A FoxPDb1 G A A C G T G C T T G T T A T T T T G A G C G G G G G A G C G C T A T C A C A T G C G G T G G A G G A G G T A G T C C A A A FoxPDa3 G A A C G T G C T T G T Y M T T T T G A G C G G G G G A G C G C T M T C A C A T G C G G T R S A G G A R G T A G T C C A R A

185 - 233

W95-182 T C A T G C C G G G T G A G G T G A A G A G C T T A G C T A C T T T A G G T G C A T C G A C T G T INDO 155 T C A T G C C G G G T G A G G T G A A G A G C T T A G C T A C T T T A G G T G C A T C G A C T G T 22615 T C C T G C C G G G T G A G G T G A G G A A C T T G G C T A C T T T A G G T G C A T C G A C T A T Fol4287 T C G C G C C G G G T G C G G T G A G G A A C T T A G C T A C T T T A G G T G C A T C G A C T G T FoxPDa2 T C A T G C C G G G T G A G G T G A A G A G C T T A G C T A C T T T A G G T G C A T C G A C T G T FoxPDb1 T C G C G C C G G G T G C G G T G A C G A A C T T A G C T A C T T T A G G T G C A T C G A C T G T FoxPDa3 T C R Y G C C G G G T G M G G T G A R G A R C T T A G C T A C T T T A G G T G C A T C G A C T G T

Figure 3.5: DNA sequence alignment of the SIX8 homologues from isolates of our study compared to homologues described in literature. Differences between isolate W95-182, INDO 155 (Foc-SIX8a) and FoxPDa2 are indicated in yellow. Differences between isolate Fol4287 and FoxPDb1 are indicated in green. Differential nucleotides between FoxPDa2, FoxPDa3 and FoxPDb1 are indicated in blue. W95-182: F. oxysporum f. sp. cubense subtropical race 4, VCG 0120 (Meldrum et al., 2012). INDO 155: F. oxysporum f. sp. cubense tropical race 4. VCG 01213, Foc-SIX8a homologue (Fraser-Smith et al., 2014). 22615: F. oxysporum f. sp. cubense subtropical race 4, VCG 0120, Foc-SIX8b homologue (Fraser-smith et al., 2014). Fol4287: F. oxysporum f. sp. lycopersici race 2 (Ma et al., 2010).

Chapter 3

3.3.4 Carbon-utilization pattern

The three pathogenic isolates belonging to St9/StQ have a similar growth profile shown by a strongly supported cluster (AU-p value = 99), characterized by a better growth on sugars, proline and threonine. The growth profile of pathogenic isolate FoxPDa3 (St1/StA) and non- pathogenic isolate FoxPDa2 (St1/StA) also clustered together with strong support (AU-p value = 99). The growth profile of the other isolates (Pathogenic isolates: Ara1, Ara2 and non-pathogenic isolate: FoxPDa9b) does not form a strongly supported cluster with other isolates (Figure 3.6).

Figure 3.6: Heatmap of the growth of different F. oxysporum isolates on minimal medium supplemented with a single carbon source (Optical density at 595nm). The standard deviation is given between brackets (n = 3). Pathogenic isolates are indicated in bold and the EF-1α and IGS sequence type is shown. The cluster dendrogram is based on Ward’s method. The numbers at the nodes give the approximately unbiased p-value, based on 1000 bootstraps.

Comparative analysis of Foc and non-pathogenic F. oxysporum populations

3.4 Discussion

In this study, evidence is shown of the introduction of pathogenicity on banana in a local F. oxysporum population in Brazil, probably through HGT.

The local F. oxysporum population obtained from the investigated farm is dominated by one group with highly similar EF-1α sequence types. The overrepresentation of a single sequence type suggests a clonal population structure, which has been observed frequently in local F. oxysporum populations (Balmas et al., 2010; Demers et al., 2015; Edel et al., 2001; Nel et al., 2006; Skovgaard et al., 2002). Pathogenic strains were only found in two different EF-1α/IGS sequence types. The main group (St9/StQ) contained no non-pathogenic isolates and several Foc isolates collected over the world have the same sequence type (MLST 25 by O’Donnell et al., 2009), suggesting that this is an introduced pathogen. However, one pathogenic strain from the farm, and two from a neighboring farm shared the same sequence type (St1/StA) with the dominant group of non-pathogenic isolates, indicating a potential exchange of pathogenicity from an individual of the St9/StQ group to a local former non- pathogenic isolate of group St1/StA.

Effector genes generally lie on a lineage specific chromosome, which contains no housekeeping genes (Ma et al., 2010). As EF-1α is a housekeeping gene, and IGS region lies between housekeeping genes, they are a part of the core genome. We have investigated the presence of putative effector genes, to get insight in the diversity of the lineage specific parts of the genome in our isolates. Of all SIX genes tested, only SIX1 was found exclusively in all pathogenic isolates, indicating its potential importance for the pathogenicity on banana. The SIX1 homologues were identical in all pathogenic isolates tested and identical to sequences of SIX1 in Australian Foc isolates of race 1, 2 and tropical race 4 (Meldrum et al., 2012; Laurence et al., 2015). The identical sequence of SIX1 among the Brazilian Foc isolates and with Foc from geographically distinct region supports the hypothesis of HGT. The sequences of Australian Foc subtropical race 4 differed by one base pair (Meldrum et al., 2012). To our knowledge, an identical sequence of SIX1 has not been found in other formae speciales.

Putative effector gene SIX8 is a multi-copy gene, which occurs in 3 to 13 copies on lineage specific and telomeric regions of Fol strains (de Sain and Rep, 2015). In the pathogenic isolates of group St9/StQ a single sequence type of SIX8 seems present. In non-pathogenic isolate FoxPDa2, which belongs to St1/StA, a combination of at least two different sequence

Chapter 3 types seems present, one similar to the sequence type found in the group St9/StQ, and a distinct one which seems to be present in a higher copy number than the SIX8 homologue found in group St9/StQ. Pathogenic isolate FoxPDa3 seems to contain more or less equal portions of the homologues found in St9/StQ group and the main homologue of FoxPDa2, supporting the suggestion that FoxPDa3 has evolved from an isolate similar to FoxPDa2 by gaining a part of the SIX8 copies of a Foc from group St9/StQ. This observation contributes to the hypothesis of HGT. However, cloning or full genome sequencing is needed to distinguish the sequence types of SIX8 and identify their copy number.

The hypothesis that FoxPDa3 has evolved from an isolate similar to FoxPDa2 is further supported by the similarity in the growth profile of FoxPDa3 and FoxPDa2. In addition, the clonal nature of the three tested pathogenic isolated belonging to St9/StQ appears to be confirmed by their similar growth profile.

Putting together the information about the population based on EF-1α and IGS, SIX1 and SIX8 homologues and the growth profile, we can suggest that isolate FoxPDa3 has gained its pathogenicity through HGT. Also for the isolates Ara1 and Ara2 this could be the case. The interplay of the exchanged pathogenicity genes with a different core genome, which is adapted to local conditions, could explain the higher virulence of isolate FoxPDa3 compared to the other Foc isolates of group St9/StQ as observed in the greenhouse assay (Chapter 2).

Topological incongruence between phylogenetic trees based on different marker genes is a strong indication of HGT (Fitzpatrick, 2011). Contradiction between the topologies of the trees based on EF-1α and IGS, which was found in this study and previously by O’Donnell et al. (2009), suggests that HGT might also take place between parts of the core genome. Vlaardingerbroek et al. (2016) observed horizontal transfer of chromosome 8 and a part of chromosome 7, both part of the core genome, between two F. oxysporum strains. HGT of both lineage specific and core chromosomes makes the evolutionary potential of FOSC complex challenging to study.

A sexual cycle has never been observed in F. oxysporum, but other mechanisms, such as transfer of active transposable elements and heterokaryosis after a parasexual cycle, can be involved in genetic change and exchange in F. oxysporum (Gordon and Martyn, 1997). Although our results support the hypothesis of HGT, it cannot be excluded that FoxPDa3 has been introduced and that the dominant non-pathogenic St1/StA population has developed from an ancestor of FoxPDa3 who has lost its pathogenicity. Complete genome sequencing of the isolates of this study is necessary to confirm the event of HGT and is an interesting

Comparative analysis of Foc and non-pathogenic F. oxysporum populations way to expand the understanding of virulence mechanisms in forma specialis cubense. Considering the limited geographic scale of this study, we cannot make assumptions on the frequency of HGT in nature. Therefore, the associated risk of dissemination of pathogenicity genes via this mechanism cannot be estimated. Ma et al. (2010) obtained HGT under specific experimental conditions with very high spore densities. It might be interesting to investigate if specific in planta condition may favor HGT.

Surprisingly, the SIX3 homologues found in this study were identical to SIX3 homologues of Fol race 3 as described in Lievens et al. (2009). Until today, they have only been found in Fol isolates. However, none of the isolates containing a SIX3 homologue caused symptoms in tomato plants. Other factors essential for pathogenicity on tomato (de Sain and Rep, 2015; Michielse et al., 2009b; Van der Does et al., 2008b) may be lacking in our isolates. Primers have been developed based on the sequence of SIX3 and SIX8 for race identification in respectively Fol and Foc (Fraser-Smith et al., 2014; Lievens et al., 2009). A SIX8 homologue highly similar to the homologue Foc-SIX8a, described as specific for Foc subtropical race 4, was found in an isolate which is non-pathogenic on banana (Fraser-Smith et al., 2014). Our results call for careful use of those primers for diagnostics without complementing pathogenicity tests. As both SIX3 and SIX8 homologues were found in isolates incapable of causing disease in banana, their role as single actor in pathogenicity on banana is doubtful.

In first instance, the isolates were collected to explore how frequently Foc could be found in non-hosts. As we were unable to isolate Foc from hosts different from banana cv. Maçã, Foc seems to colonize non-hosts rather seldomly. In a study conducted in an Australian banana plantation only 6 out of 154 isolates from weeds were identified as Foc tropical race 4 (Hennesey et al., 2005). On the contrary, studies on other formae speciales report isolation frequencies in the range of 50% to 94% of all isolates collected from non-hosts (Altinok, 2013; Elmer and Lacy, 1986; Gordon et al., 1989; Helbig and Carroll, 1984; Scott et al., 2014). The sample size of our study does not allow making firm conclusions on capacity and frequency of Foc to colonize non-hosts and more extensive sampling is necessary.

Although banana is an introduced crop in Brazil, pathogenic isolates with marker genes identical to the main local non-pathogenic population have been found. This suggests that Foc populations are dynamic and from introduced pathogens, local pathogenic strains can evolve, probably through HGT. Through sequencing of two marker genes of the core genome and putative effector genes, we could identify isolates that probably received pathogenicity genes via HGT. Full genome sequencing of those isolates is necessary to confirm HGT and is an excellent opportunity to further investigate genes involved in virulence

Chapter 3 on banana. Of the putative effector genes, only SIX1 seems to be linked with the disease on banana, as SIX3 and SIX8 homologues were also found in non-pathogenic isolates.

3.5 Acknowledgements

The authors would like to thank the farmers Geraldo Maçela dos Santos, Sebastião Elias dos Santos, Geraldo Cândido da Silva and Maria Aparecida de Almeida Pedrosa for their hospitality, Irene Cardoso and Ivo Jucksch to facilitate the collection and Filip Snauwaert for his technical assistance. This study was funded by a PhD grant of the Scientific Research- Flanders (FWO) to P. D.

Chapter 4

Interaction between Fusarium oxysporum f. sp. cubense and non-pathogenic Fusarium oxysporum in graminoids

Authors Pauline Deltoura, Soraya C. Françaa, Monica Höftea

Affiliations aLaboratory of Phytopathology, Department of Crop Protection, Faculty of Bioscience Engineering, Ghent University, Coupure Links 653, 9000, Ghent, Belgium

Chapter 4

Abstract

Fusarium oxysporum is a species complex which comprises many wilt pathogens, such as Fusarium oxysporum f. sp. cubense (Foc). Foc causes Fusarium wilt of banana, also known as Panama disease. Moreover, the F. oxysporum species complex contains many non- pathogens, some of which can protect plants against diseases. F. oxysporum has been reported as one of the most common endophytes of graminoids. Yet, the endophytic associations have been poorly characterized. Many graminoids commonly present in banana-based farming systems as weeds, cover crops or intercrops can be potential symptomless carriers of Foc. Surprisingly, none of the endophytic F. oxysporum isolates we have obtained from graminoids growing in a banana-agroforestry system in Brazil, resulted to be pathogenic to banana. This finding led us to investigate the capacity and competitiveness of a pathogenic isolate Foc race 1 (FocR1) and three non-pathogenic F. oxysporum isolates to colonize two graminoids: Brachiaria decumbens, a common cover crop, and Cyperus iria, a weed. Autoclaved soil was inoculated with either a single isolate or a combination of one non-pathogenic isolate with the pathogenic isolate FocR1. Root and shoot colonization was quantified by plating and qPCR. We found that all isolates colonized the roots of both graminoids to the same extent when inoculated separately. However, with combined inoculations the root colonization by the pathogenic isolate FocR1 was consistently lower (6 to 115 times) than that by the non-pathogen. The reduced competitiveness of FocR1 could not be explained by its lower growth on different carbon sources. This study shows that interactions among members of the F. oxysporum complex may modulate the impact of symptomless hosts on the pathogen inoculum population. The results suggest that graminoids are more likely to act as reservoir of non-pathogenic F. oxysporum, than of Foc, by which the pathogen inoculum potential can decrease. Further research is needed to investigate if reduced competitiveness on symptomless hosts is common to Foc isolates.

Interaction of Foc and non-pathogenic F.oxysporum in graminoids

4.1 Introduction

Fusarium oxysporum is a diverse species complex, best known for its wilt causing pathogens. Pathogenic F. oxysporum isolates are generally host specific and subdivided in formae speciales accordingly. However, many F. oxysporum isolates are endophytes of plants without causing symptoms and are considered to be putative non-pathogens. Some non-pathogenic F. oxysporum isolates can act as biocontrol agents against pathogens and nematodes (Alabouvette et al., 2009).

The management of Fusarium wilt is challenging because of the long term survival of the pathogen in soil in absence of a susceptible host plant. F. oxysporum can persist in soil through survival structures, chlamydospores and/or sclerotia, by saprophytic growth on organic material in soil or via the ability to colonize plants in which they do not cause symptoms (Ploetz, 2015).

Although F. oxysporum is one of the most commonly encountered endophytic fungi of plants, the endophytic association is poorly characterized (Malcolm et al., 2013). Some authors have shown that the endophytic F. oxysporum population associated with different tomato varieties or plant species is specific and different from the surrounding soil (Demers et al., 2015; Edel et al., 1997). Leoni et al. (2013) showed that F. oxysporum f. sp. cepae colonized roots of different plant species to a different extent. Further study on multiple endophytic associations is required to understand the role of plants as symptomless carriers in the persistence and inoculum increase of pathogenic F. oxysporum in soil. Particularly the role of graminoids seems to be important since several studies have shown that graminoids are good symptomless hosts of F. oxysporum (Elmer and Lacy, 1987; Gordon et al., 1989; Gordon and Martyn, 1997; Katan, 1971).

F. oxysporum f. sp. cubense (Foc) is the causal agent of Fusarium wilt on banana. Besides bananas, many other plants can be found in banana-based cropping systems, such as weeds, cover crops and intercrops, which can serve a symptomless carriers of Foc. Cover crops are selected based on many criteria, such as good competition for weeds, ability to cover the soil, minimize the competition for water and nitrogen, tolerance for shade, but their impact on Fusarium wilt is overlooked since little is known about the impact of symptomless hosts on inoculum levels of Foc (Tixier et al., 2010). In a previous study in a banana- agroforestry system in Brazil, several graminoids were sampled to explore the host range of

Chapter 4

Foc. Although F. oxysporum could be isolated from half of the graminoids, none of the isolates appeared to be pathogenic on banana (Chapter 3). Also Hennessy et al. (2005) found that in a banana plantation in Australia only 3.75% of all F. oxysporum isolates collected from weeds were pathogenic on banana.

Based on the low isolation frequency of Foc in our previous study and the observations of Hennessy et al. (2005) we performed a study to test whether (1) Foc has a lower capability to colonize symptomless host plants in comparison to non-pathogenic isolates, or (2) Foc has lower competitiveness to colonize symptomless hosts than non-pathogenic F. oxysporum. Two different graminoids, Brachiaria decumben, a common cover crop and Cyperus iria, a weed, were used as model plants to consider the plant species specificity of the colonization. As an additional objective, the isolate and plant specificity of the colonization of areal plant parts were tested. Insight in the colonization characteristics of several F. oxysporum isolates can aid on evaluating the role of symptomless hosts on the disease.

4.2 Material and methods

4.2.1 Fungal isolates and inoculum preparation

The isolates were obtained from a farm in Pedra Dourada, MG, Brazil (Table 4.1). Their pathogenicity to banana cultivar Silk has been tested before (Chapter 3). Isolates that caused no symptoms on banana are called non-pathogenic F. oxysporum (np-Fox) throughout the manuscript.

A chlamydospore-rich inoculum was prepared according to the method of Smith and Snyder (1971). Air-dried soil was autoclaved for 2 consecutive days and subsequently inoculated with a conidiospore suspension of an isolate grown for 14 days on Potato Dextrose Agar (PDA) (107 spores g-1 soil) and incubated for 3 weeks at room temperature. Final inoculum level, checked by dilution plating on PDA supplemented with 100 mg l-1 streptomycin, was in the range of 105 cfu g-1 soil.

Interaction of Foc and non-pathogenic F.oxysporum in graminoids

Table 4.1: F. oxysporum isolates used in this study

Working Official Origin Pathogenicityb St EFc St IGS c Name Namea FocR1 FoxPDa15 Banana cv. Maçã f. sp. cubense race 1 9 Q npM MrPDb1 Banana cv. Maçã np 11 W npC GrPDa1b Cyperus iria np 1 A npD GsPDb1c Digitaria insularis np 12 O a Isolate name as listed in Chapter 3 b np: non-pathogenic on banana cv. Silk c St (sequence type) of EF-1α and IGS as defined in Chapter 3

4.2.2 Colonization of C. iria and B. decumbens by F. oxysporum isolates

Two pot experiments were established in growth chambers at 26°C, 16/8 day/night light regime to evaluate the capacity and competitiveness of Foc to colonize roots and shoots of graminoids in comparison to non-pathogenic isolates. Two different graminoids species were tested: Brachiaria decumbens, a common cover crop and Cyperus iria, a weed. Seeds of B. decumbens (supplied by Oswaldo Rodriques Milagres Júnior) and C. iria (Herbiseed, Reading, UK) were surface sterilized by submerging 1 min in 70% ethanol, 15 min in 1% sodium hypochlorite (NaOCl) and washing three times with sterile water. The seeds were pregerminated on sterile filter paper.

In the first explorative experiment, one week old seedlings were planted in autoclaved substrate (50:50 sand/potting soil) inoculated with chlamydospore-rich inoculum. The treatments consisted of separate inoculation with the pathogenic isolate FocR1, or the non- pathogenic isolates npC or npD in a concentration of 104 cfu g-1 substrate. In addition, seedlings were planted in pots inoculated with the combination of FocR1 (104 cfu g-1 substrate) and of one of the non-pathogenic isolates npC or npD (104 cfu g-1 substrate). As control, plants were grown in non-inoculated substrate. Four repetitions with two plants per pot were used. After 8 weeks, the plants were carefully uprooted and evaluated for wilting and internal discoloration. Roots and shoots were carefully washed and surface sterilized (2 min in 1% NaOCl). Per plant 10 root and 5 shoot pieces of 1 cm were plated on PDA supplemented with 100 mg l-1 streptomycin. Fungal growth was observed during 1 week and isolates were distinguished based on distinct morphological appearance (Figure 4.1).

In the second experiment, a slightly modified setup was used. Sandy soil collected in Denderbelle, Belgium, was autoclaved for 1 h during 3 consecutive days. Seedlings of B. decumbens and C. iria were planted in sterilized soil with different treatments. Firstly, the

Chapter 4 sterilized soil was inoculated with a single F. oxysporum isolate (FocR1, npC, npD or npM) in the concentration of 5.103 cfu g-1 soil. In addition, combined inoculations were performed with the pathogen FocR1 (5.103 cfu g-1 soil) and one of the non-pathogenic isolates npM, npC or npD (5.103 cfu g-1 soil). Finally, in one treatment the soils were inoculated with a combination of all four isolates (5.103 cfu g-1 soil for each isolate). Control plants were grown in non- inoculated soil. Four repetitions (pots) with two plants per pot were used. Plantlets were grown during 11 weeks before evaluation of wilt symptoms and internal discoloration. All plants were uprooted and roots were carefully washed to remove all residual soil. Two plants were cut into pieces to be plated as described before. Roots and shoots of the two other plants were kept on -20°C for DNA extraction. One soil sample was used for dilution plating on selective Komada medium (Komada, 1975) and another kept at -20°C before DNA extraction. To investigate if shoot colonization of C. iria is isolate specific, the inoculation with single isolates was repeated on C. iria.

Figure 4.1: Example of morphological recognition of F. oxysporum isolates used in this study, growing on PDA with streptomycin.

Interaction of Foc and non-pathogenic F.oxysporum in graminoids

4.2.3 Primer design and qPCR

Specific primers were designed from SNP’s in the translation elongation factor-1α (EF-1α) sequence that were discriminative for the isolates FocR1, npM, npC or npD (Table 4.2). Primers were manually selected and evaluated by using the program Oligo Analyzer (Gene Link) for primer-dimer formation and self-annealing. The specificity of the selected primers was first tested by conventional PCR, and the annealing temperature was adjusted. The PCR mixture contained 5 μl buffer, 0.5 μl dNTP’s 0.15 μl Taq, 1 μl of each primer, 1 μl DNA sample and 16.35 μl MilliQ water. The PCR program was as follows: 5 min on 95°C set-up denaturation, followed by 30 cycles of 1 min denaturation on 95°C, 1 min annealing on 58- 67°C and 3 min extension on 72°C, concluded with a final extension of 7 min on 72°C.

For qPCR, 20 μl reactions, containing 2 μl template, 300 nM of each primer, 10 μl SYBR Green Mastermix (Stratagene/bioconnect) and 6 µl DNAfree water, were used. The thermocycling profile consisted of a first cycle of 10 min at 95°C, 40 cycles of 30 s at 95°C and 1 min at 64°C or 67°C (respectively for the three primer pairs PathF/PathR; S3F/S3R and Gr1_356F/Gr1_356R and the primer pair Gr3F/Gr3R) and a last cycle of 1 min 95°C, 30 s at 64°C or 67°C (respectively) and 30 s at 95°C. PCR amplifications were conducted with the Mx3005P real time PCR detection system (Stratagene). Fluorescence data were collected at the end of each elongation phase. All primer pairs resulted in single fragments based on the melting curve analysis. Linear DNA standard curves for qPCR assays were obtained from 10 fold serial dilution of genomic DNA of the appropriate isolate (10-1 pg μl-1 to 103 pg μl-1). Efficiency and sensitivity were calculated (Table 4.2). The limit of quantification was determined as the lowest dilution for which the standard curve gave consistent results. As the primers were developed to be used in ecological studies, a cutoff was installed at the analytical sensitivity of the standard curve (Caraguel et al., 2011), which was corresponding to a concentration of 1 pg µl-1 pure fungal DNA for the primer pairs PathF/PathR, S3F/S3R, Gr1_356F/Gr1_356R and 5 pg µl-1 for Gr3F/Gr3R. The corresponding Ct value was 34 for all primer pairs, which was set as limit of quantification. Subsequently, specificity of the qPCR reaction of all four primer pairs was analyzed using 1 ng μl-1 genomic DNA of the isolates FocR1, npM, npC and npD. Above the limit of quantification, there was no nonspecific detection (Table 4.3). The specificity was not tested on other fungi, as the primers were designed to distinguish the different F. oxysporum isolates in controlled environment.

Chapter 4

After crushing the shoots and roots in liquid nitrogen, DNA was extracted from a 70-90 mg sample with the Invisorb spin plant mini kit (Stratec Molecular, Berlin). DNA from soil was extracted from a 250 mg sample with Mo Bio PowerSoil DNA isolation kit, following the manufacturer’s instructions. The qPCR reaction was performed as described above and quantification was done by calibrating to a standard curve with a 10 fold dilution series of genomic DNA in sterile water (1 to 103 pg μl-1 (FocR1, npM, npC)). For primer pair Gr3F/Gr3R, 5 pg μl-1 was used instead of 1 pg μl-1, as 1 pg μl-1 fell beyond the analytical sensitivity. All samples were run in duplicate. Ct values which fell beyond the limit of quantification were maintained in the data processing as they display the best estimate for data beyond the analytical sensitivity (Caraguel et al., 2011; Hewett and Ganser, 2007). Ct values of 40 were used as substitution for data where no Ct value was obtained to count for the most optimistic estimate (Hewett and Ganser, 2007). The substitution method for censored data delivers the smallest bias for small sample sizes (Hewett and Ganser, 2007).

Table 4.2: Primers used in this study, standard curve, linearity (R2) and amplification efficiency

Name Primer Sequence (5’ → 3’) Amplicon Standard curve (x in fg) R2 Amplifi- target name size (bp) cation isolate efficiency FocR1 PathF CGACAATGAGCATATCTGCCATT 341 Y= -3.383 x + 44.944 0.9904 97.51 PathR CATCGAGGTTGTGAGAATGGA npM S3F CTCACAACCTCAATGAGTGT 239 Y= -3.7746 x + 46.293 0.9902 84.05 S3R CGAAGAGAAGTAGAATGAAGCAA npC Gr1_356F TTACCCCGCCACTTGAGCGAA 103 Y= -3.6963 x + 46.995 0.996 86.44 Gr1_356R GTTGAATTGTTAGTGACTGCTTC npD Gr3F CGAGAAGTTCGAGAAGGTTAGTT 231 Y= -3.8217 x + 48.712 0.9937 82.69 Gr3R TGCTTGACACGTGACGATGCG

Table 4.3: Ct values with 1 ng μl-1 gDNA of each isolate

Target Primers FocR1 npM npC npD PathF/PathR 23.58 S3F/S3R 22.84 Gr1_356F/R 35.84 37.37 23.74 Gr3R/F 35.62 24.78

Interaction of Foc and non-pathogenic F.oxysporum in graminoids

On the plant samples which were inoculated by one single isolate or not inoculated, qPCR was performed with all primer pairs to check for sensitivity, specificity and accuracy in environmental samples. The sensitivity was calculated as: r = [TP/ (TP + FN)]. The specificity was calculated as s = [TN/(TN + FP)]; and the accuracy as a = [(TP + TN)/ all samples]. TN = true negative; FP: true positive; FP: false positive (Gómez-Doñate et al., 2016). The analytical cutoff was used to determine whether samples were negative or positive. All true positive samples were detected, resulting in 100% sensitivity, and the accuracy was between 97.5% and 87.5% for all primer pairs (Table 4.4).

Table 4.4: Limit of quantification (LOQ), Ct value at LOQ, sensitivity, specificity and accuracy determined on the root colonization of C.iria and B. decumbens by separate isolates.

Primer LOQ (ng) cutoff Sensitivity Specificity Accuracy PathF/PathR 0.001 34 1 31/32 39/40 S3F/S3R 0.001 34 1 28/32 36/40 Gr1_356F/R 0.001 34 1 31/32 39/40 Gr3R/F 0.005 34 1 31/32 39/40

4.2.4 Data analysis

In the result section, the data of the single inoculation of the two experiments are presented under one paragraph, as both deal with the intrinsic capacity of F. oxysporum to colonize the roots of the plant. Data obtained from plating and qPCR are shown in the same paragraph to demonstrate their complementary. Differences in the level of colonization were statistically tested with a Kruskal Wallis test, both for plating and qPCR data. Also shoot colonization levels, as determined with qPCR, have been tested with a Kruskal Wallis test.

The root colonization after combined inoculations of the two experiments is also shown under one section. Within the same sample, the quantity in the roots of the different isolates as determined with qPCR was compared by a paired t-test. To compare the colonization level of one isolate in combined inoculation with the level when inoculated alone, a Wilcoxon rank- sum test was used. The population level of the different isolates in soil after combined inoculation was determined by plating and compared with a paired t-test. Table 4.5 gives an overview of the experiments and the data shown in the different paragraphs.

Chapter 4

Table 4.5: overview of the experiments Experiment 1 Experiment 2 Extra (B. decumbens + (B. decumbens + (C. iria) C. iria) C. iria) Paragraph Inoculated Method Inoculated Method Inoculated Method with with with 4.3.1.1 Root • FocR1 Plating • FocR1 Plating colonization after single • npC • npC + qPCR inoculation • npD • npD • npM

4.3.1.2 Shoot • FocR1 qPCR • FocR1 qPCR colonization after single • npC • npC inoculation • npD • npD • npM • npM

4.3.1.3 Root • FocR1+npC Plating • FocR1+npC Plating colonization after • FocR1+npD • FocR1+npD + qPCR combined inoculation • FocR1+npM • FocR1+npC+ npD+npM

4.3.1.4 Soil • FocR1+npC Plating colonization dynamics • FocR1+npD after combined • FocR1+npM inoculation • FocR1+npC+ npD+npM

4.2.5 Phenotyping of F. oxysporum isolates

A Biolog FF MicroPlate (BIOLOG, Hayward, CA, USA) was used to investigate carbon source utilization by the pathogenic and non-pathogenic isolates, following the protocol adapted from Michielse et al. (2009b). Each well was filled with 150 μl of a conidial suspension resulting in 104 spores per well. The plates were incubated at 25°C and the turbidity was measured at 750 nm with a multimode microplate reader (Infinite M200 PRO NanoQuant, Tecan). Three repetitions were used per isolate. The average well turbidity was calculated after subtracting the turbidity of the water control from each value and tested with Tukey's HSD (honest significant difference) test.

Interaction of Foc and non-pathogenic F.oxysporum in graminoids

4.3 Results

4.3.1 Colonization by F. oxysporum isolates

4.3.1.1 Root colonization after single inoculation

Seedlings of B. decumbens and C. iria were planted in sterilized soil inoculated with the pathogenic isolate FocR1 or the non-pathogenic isolates npC, npD (experiment 1 and 2) or npM (experiment 2). Root colonization was analyzed by plating (experiment 1 and 2) and by qPCR (experiment 2) after 8 weeks of growth. Root colonization analysed by plating was not different between isolates or plants in both experiments (Figure 4.2). Like observed by plating, qPCR quantification of the root colonization revealed no big differences between the isolates (Figure 4.3). All samples were above the detection limit of quantification. None of the plants showed vascular discoloration or wilt symptoms.

A B a a ab a a 100 ab b ab

100 a a b 90 bc

(%) 90 bcd 80 80 a 70 70 60 cd 60 FocR1 50

F. oxysporum oxysporum F. 50 npM 40 40 30 30 d npC b 20 20 npD 10 b 10 1.2 0 0.81 0 0.40.6

0.20

Ctrl Ctrl Ctrl Ctrl

npC npC npC npC

npD npD npD npD

npM npM

FocR1 FocR1 FocR1

FocR1 1 Root pieces colonized by colonized pieces Root C. iria B. decumbens C. iria B. decumbens

Figure 4.2: Root colonization by different F. oxysporum isolates after single inoculation, analyzed by plating on PDA with streptomycin. A: First experiment, B: Second experiment. Different letters show statistical differences (Kruskal Wallis, Bonferroni, p = 0.05). Bars show the standard deviation (n = 4).

Chapter 4

C. iria B. decumbens 7 a a 7 a a a

6 a 6 a a

root root

1 -

5 5 4 4 3 3 2 2

fresh weight) fresh 1 1 0 0

Log (pg fungal DNA g DNA fungal Log(pg npM npC npD FocR1 npM npC npD FocR1

Figure 4.3: Root colonization by different F. oxypsorum isolates after single inoculation in the second experiment measured by qPCR. The error flags show the standard deviation. All values were above the limit of quantification (4/4; n = 4). Different letters indicate significant differences (Kruskal Wallis, Bonferroni, p = 0.05).

4.3.1.2 Shoot colonization after single inoculation

The plants that were analysed for root colonization by qPCR, were also analysed for shoot colonization. In B. decumbens, only in two samples F. oxysporum could be detected above the limit of quantification. On the contrary, in C. iria the non-pathogenic isolates npM and npD were detected in at least three of the four samples, while isolate npC and FocR1 were below the detection threshold (Figure 4.4). However, the isolate specific colonization of C. iria could not be confirmed in the repetition since in this experiment FocR1 was detected in all samples, besides isolates npM and npD. Only the colonization by isolate npC remained low (Figure 4.4). Results of plating are not shown as contamination hindered accurate counts.

Interaction of Foc and non-pathogenic F.oxysporum in graminoids

6 C. iria (1) 6 C. iria (2) 6 B. decumbens 5 5 5

4 4 4

shoot fresh shoot

- 3 3 3 weight) 2 2 2

1 1 1 Log (pg DNA g DNA (pg Log 0 0 0 npM npC npD FocR1 npM npC npD FocR1 npM npC npD FocR1

npM npC npD FocR1 npM npC npD FocR1 npM npC npD FocR1 2.85 1.00 3.57 1.10 3.70 2.53 4.10 4.18 1.87 1.52 2.40 1.57 (0.81) (0.49) (0.96) (0.33) (1.17) (0.91) (1.07) (0.77) (1.10) (0.26) (0.17) (1.28) a b a b a' a' a' a' a” a” a” a”

Figure 4.4: Shoot colonization of C. iria and B. decumbens by different F. oxysporum isolates detected by qPCR. C. iria (1) and B. decumbens show the results of plants of the second experiment (n = 4). The shoot colonization of C. iria was also analysed in a partial repetition of experiment 2, indicated as C. iria (2). Different data points are shown with diamonds. Data points falling below the limit of quantification are indicated with a lighter colour. In the table, mean values are shown with the standard deviation between brackets. Within rows, statistically significant differences are indicated with a different letter (Kruskal Wallis, Bonferroni, p = 0.05).

4.3.1.3 Root colonization after combined inoculation

When analyzing root colonization by plating after combined inoculation with Foc and one non-pathogenic isolate, only from a small fraction of root fragments Foc isolate FocR1 could be detected (Figure 4.5). Analysis of the samples of the second experiment by qPCR confirmed these findings. Upon co-inoculation, pathogenic isolate FocR1 consistently colonized the roots of B. decumbens and C. iria at lower densities than non-pathogenic isolates (from 6 to 115 times lower), clearly demonstrated by the paired statistics shown in Figure 4.6.

However, isolate specific results were found when the level of root colonization after combined inoculation was compared with single inoculation. In the combined inoculation, the level of root colonization by isolate npM remained the same as that of single inoculation, while FocR1 colonization was significantly decreased. More so, combining with isolate npC resulted in a lower root colonization by pathogenic isolate FocR1 in comparison to single inoculation. In contrast to isolate npM, root colonization by isolate npC decreased after combining with Foc, although, not significantly in the roots of B. decumbens. After combined inoculation, an increased level in the root colonization was observed for isolate npD, which

Chapter 4 was significant in B. decumbens. Compared to the single inoculation, there was no significant decrease in the root colonization of isolate FocR1 after combination with npD (Figure 4.6). In all combinations, no plant specific effect of root colonization could be observed.

When all four isolates were co-inoculated, pathogenic isolate FocR1 could only be detected in one of the four samples of each plant species. Moreover, the root colonization by npC was significantly reduced compared to single inoculation, while a large variation was found in the root colonization by isolate npM. Only isolate npD maintained the same level of root colonization as with single inoculation. In both plant species, root colonization was similar (Figure 4.7).

A B

100 100

(%) 90 90 80 80 70 70 FocR1 ab F. oxyporum oxyporum F. 60 60 npM 50 50 a npC ab a ab 40 ab 40 ab ab npD bc 30 30 no F. oxysporum b 1.2 20 20 0.81 0.40.6 0.20 cd 10 10 de 1 Root pieces colonized by colonized pieces Root c c c c de de de 0 0 e npC npD npC npD npM npC npD npM npC npD C. iria B. decumbens C. iria B. decumbens

Figure 4.5: Root colonization with combined inoculation of the pathogenic isolate FocR1 with the non-pathogenic isolates npM, npC or npD. The evaluation was done by morphological recognition on PDA with streptomycin. A: The first experiment, B: The second experiment. Proportions indicated with different letters are significantly different (Kruskal Wallis. Bonferroni, p = 0.05, n = 4)

Interaction of Foc and non-pathogenic F.oxysporum in graminoids

C. iria

7 *** 7 * 7 * p = 4.02 10-5 p = 0.024 p = 0.0314

6 6 6

root

1 - 5 5 5 4 4 4 3 3 3

freshweight) 2 2 2

1 1 1 Log (pg Log fungal DNA g 0 0 0 npM1 FocR12 npC1 FocR12 npD1 FocR12

Inoculation npM FocR1 npC FocR1 npD FocR1 Single 5.55 (0.51) a 4.90 (0.57) a 5.67 (0.25) a 4.90 (0.57) a 5.69 (0.24) a 4.90 (0.57) a Combined 5.24 (0.30) a 3.34 (0.67) b 4.57 (0.63) b 3.17 (1.16) b 6.23 (0.60) a 4.46 (1.43) a

B. decumbens 7 7 * ** p = 0.037 7 p = 0.005 p = 0.064

root 6 6

1 6

5 5 5 4 4 4 3 3 3

freshwheight) 2 2 2 1 1 1 Log( pg Log( fungal DNA g 0 0 0 npM1 FocR12 npC1 FocR12 npD1 FocR12

Inoculation npM FocR1 npC FocR1 npD FocR1 Single 5.04 (0.70) a 5.16 (0.35) a 5.63 (0.20) a 5.16 (0.35) a 5.02 (0.31) b 5.15 (0.35) a Combined 5.27 (0.36) a 3.58 (0.36) b 5.14 (0.54) a 3.63 (1.06) b 5.74 (0.49) a 3.83 (1.74) a

Figure 4.6: Root colonization by the pathogenic isolate FocR1 and the non-pathogenic F. oxysporum isolates npM, npC, npD with combined inoculation in the second experiment, analysed by qPCR. Separate data points are indicated with a diamond. Data of the same plant sample are connected with a dashed line. Data points falling below the limit of quantification are indicated with a lighter colour. Differences between the pathogenic and non-pathogenic isolates were tested with a paired t-test (p value indicated in the figures; * p < 0.05; ** p < 0.01; *** p < 0.001). Mean values in combined and single inoculation are compared with Wilcoxon rank-sum test and shown in the tables below the figures. Within columns, different letters indicate significant differences (p = 0.05).

Chapter 4

7 C. iria 7 B. decumbens 6 6

5 5

rootfresh

1 - 4 4

3 3 weight) 2 2 1 1 Log (pg Log DNA g 0 0 1 2 3 4 npM1 npC2 npD3 FocR14 npM npC npD FocR1

Inoculation npM npC npD FocR1 npM npC npD FocR1 Single 5.54 5.67 5.69 4.90 5.03 5.63 5.02 5.16 (0.51) a (0.25) a (0.24) a (0.57) a (0.70) a (0.20) a (0.31) a (0.35) a Combined 3.55 4.29 5.22 2.08 4.00 4.05 4.92 2.77 (1.45) a (0.57) b (0.57) a (1.26) b (1.52) a (0.33) b (0.22) a (0.89) b

Figure 4.7: Root colonization after combined inoculation of four F. oxysporum isolates: pathogen FocR1 and the non-pathogenic isolates npM, npC, npD. Separate data points are indicated with a diamond. Data of the same sample are connected with a dashed line. Data points falling below the limit of quantification are indicated with a lighter colour. Differences between the different isolates were tested with a paired t-test. Only in C. iria the level of npM and npC were not significantly different. Mean values in combined and separate inoculation are compared with Wilcoxon rank-sum test and shown in the tables below the figures. Within columns, different letters indicate significant differences (p = 0.05).

4.3.1.4 Soil colonization dynamics after combined inoculation

The different root colonization could be an effect of differences in soil colonization. Therefore, we investigated the population level of each isolate in soil in the combined inoculations at the end of the experiment. Based on plating on selective medium, the population levels in soil were variable in most cases and no general trend was observed. Only in the soil after growth of B. decumbens, non-pathogenic isolates npM and npD were more abundant than pathogenic isolate FocR1 (Figure 4.8). In combination of all four isolates after growth of B. decumbens, the level of npD was significantly higher than the level of FocR1 in soil (Figure 4.9). Analysis by qPCR showed small, non-significant differences in the level of the isolates in soil. Only the difference between FocR1 and npD after growth of B. decumbens was significant (Supplementary material S4.1).

Interaction of Foc and non-pathogenic F.oxysporum in graminoids

C. iria

4 p = 0.333 4 p = 0.173 4 p = 0.540 3 3

soil) 3

1 - 2 2 2

1 1 1 Log (cfu Log g 0 0 0 1 2 1 2 1 2 npM FocR1 npC FocR1 npD FocR1

B. decumbens 4 p = 0.031*

4 p = 0.047* 4 p = 0.192

3 3 3

soil)

1 - 2 2 2

1 1 1 Log (cfu Log g 0 0 0 npM1 FocR12 npC1 FocR12 npD1 FocR12

Figure 4.8: Soil populations of the pathogenic isolate FocR1 and the non-pathogenic F. oxysporum isolates npM, npC, npD after growth of C. iria or B. decumbens. Soil samples were collected from the combined inoculation in the second experiment and analysed by plating. Separate data points are indicated with a diamond. Data of the same soil sample are connected with a dashed line. Soil samples where no colony was found are indicated with a lighter colour and pictured as zero values. Differences between the pathogenic and non-pathogenic isolates were tested with a paired t-test (p value indicated in the figures).

Chapter 4

C. iria B. decumbens

4 4 *

3 3

soil)

1 - 2 2

1 1 Log (cfu Log g

0 0 npM1 npC2 npD3 FocR14 npM1 npC2 npD3 FocR14

Figure 4.9: Soil populations of the pathogen FocR1 and the non-pathogenic isolates npM, npC, npD after combined inoculation and growth of C. iria or B. decumbens , analysed by plating. Separate data points are indicated with a diamond. Data points of the same soil sample are connected with a dashed line. Soil samples where no colony was found are indicated with a lighter colour and pictured as zero- values. Differences between the different isolates were tested with a paired t-test and significant differences (p = 0.05) are indicated with an asterisk.

4.3.2 Phenotyping of the F. oxysporum isolates

To investigate the possible involvement of specific growth capacity of the F. oxysporum isolates in root colonization competitiveness, the isolates were grown on 95 different carbon sources. Only non-pathogenic isolate npC had a higher average growth level than pathogenic isolate FocR1. The two other non-pathogenic isolates, npM and npD, had a lower average growth and lower maximum growth in comparison with isolate FocR1. Isolate npD, which demonstrated the strongest root colonization in the combined inoculation, had the lowest average growth of all isolates (Figure 4.10). Carbon sources where little growth was observed were generally very similar among the isolates, but the main sources that sustained good growth differed. Pathogenic isolate FocR1 grew especially well on L-proline, L-pyroglutamic acid and D-cellobiose in comparison to the other isolates. All isolates grew well on D-trehalose (Figure 4.10 and Figure 4.11).

Interaction of Foc and non-pathogenic F.oxysporum in graminoids

A 0.7 a 0.6 b 0.5

0.4 c d 0.3

0.2 Absorbance 750 nm 750 Absorbance 0.1 0 npM npC npD FocR1

B 35 35 35 35 30 30 30 30 25 25 25 25 20 20 20 20 15 15 15 15 10 10 10 10 5 5 5 5

0 0 0 0

Number of carbon sources carbon of Number

1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9

0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7

------

< 0.1 <

< 0.1 < < 0.1 <

< 0.1 < OD OD OD OD

0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5

0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 OD-value C

npM npC npD FocR1 1 Stachyose Turanose D-Trehalose D-Trehalose 2 Amygdalin D-Mannitol Turanose D-Xylose 3 D-Xylose L-Ornithine β-Methyl-D-Glucoside L-Proline 4 Maltotriose Maltose Maltotriose L-Pyroglutamic Acid 5 D-Trehalose D-Trehalose D-Raffinose N-Acetyl-D-Glucosamine 6 α-Methyl-D-Galactoside D-Galactose Quinic Acid D-Cellobiose 7 Maltose D-Xylose L-Glutamic Acid Quinic Acid 8 N-Acetyl-D-Glucosamine D-Fructose N-Acetyl-D-Glucosamine Maltotriose 9 Quinic Acid α-Methyl-D-Galactoside Dextrin Gentiobiose 10 D-Melezitose β-Methyl-D-Glucoside Maltose L-Glutamic Acid

Figure 4.10: Phenotyping of the pathogen FocR1 and the non-pathogenic isolates npM, npC, npD. A: Average well turbidity of all 95 carbon sources of the Biolog FF plate. Statistically different averages are indicated with different letters (Tukey’s HSD, p = 0.05, n = 3). The error bars show the standard deviation. B: Histogram showing the distribution of the optical density (OD) values of the 95 carbon sources. C: The 10 carbon sources on which most dense growth was measured per isolate, ranked from highest OD to lowest.

Chapter 4

npM npC npD FocR1 npM npC npD FocR1 0.00 0.01 0.00 0.00 N-Acetyl-L-Glutamic Acid 0.57 0.64 0.33 0.39 Sucrose 0.00 0.01 0.00 0.01 Uridine 0.11 0.32 0.11 0.39 L-Threonine 0.01 0.01 0.00 0.01 Adenosine-5'-Monophosphate 0.52 0.35 0.44 0.43 Tween 80 0.00 0.00 0.00 0.01 N-Acetyl-D-Galactosamine 0.24 0.88 0.26 0.43 L-Rhamnose 0.01 0.00 0.01 0.01 N-Acetyl-D-Mannosamine 0.84 0.59 0.27 0.45 Amygdalin 0.02 0.01 0.00 0.01 Glycyl-L-Glutamic Acid 0.39 0.67 0.30 0.45 D-Mannose 0.02 0.02 0.02 0.02 α-Cyclodextrin 0.44 0.43 0.37 0.45 Salicin 0.01 0.03 0.00 0.02 i-Erythritol 0.86 1.02 0.48 0.45 Stachyose 0.01 0.07 0.00 0.04 α-D-Lactose 0.31 0.54 0.36 0.47 Succinic Acid 0.01 0.04 0.03 0.05 Alaninamide 0.17 0.68 0.15 0.47 Putrescine 0.07 0.13 0.07 0.05 γ-Hydroxy-butyric Acid 0.31 0.98 0.14 0.48 D-Arabitol 0.02 0.08 0.02 0.05 Xylitol 0.24 1.27 0.43 0.48 m-Inositol 0.03 0.04 0.05 0.05 Glucose-1-Phosphate 0.40 0.81 0.35 0.48 β-Methyl-D-Galactoside

0.03 0.02 0.04 0.05 Glucuronamide 0.45 0.50 0.53 0.51 Dextrin 0.08 0.07 0.12 0.06 β-Cyclodextrin 0.47 0.35 0.31 0.51 Glycogen 0.05 0.75 0.06 0.06 L-Sorbose 0.51 0.75 0.56 0.57 D-Raffinose 0.02 0.06 0.00 0.07 Lactulose 0.37 1.58 0.38 0.60 D-Mannitol 0.00 0.12 0.00 0.08 Sedoheptulosan 0.46 0.89 0.40 0.60 Fumaric acid 0.04 0.08 0.05 0.09 D-Arabinose 0.43 0.98 0.43 0.64 D-Glucuronic Acid 0.10 0.16 0.08 0.09 L-Fuctose 0.45 0.69 0.15 0.66 D-Ribose 0.09 0.15 0.11 0.10 L-Phenylalanine 0.35 0.75 0.46 0.66 L-Aspartic Acid 0.05 0.11 0.01 0.10 D-Tagatose 0.63 1.01 0.37 0.67 L-Arabinose 0.08 0.12 0.04 0.12 α-Methyl-D-Glucoside 0.26 0.40 0.47 0.68 p-Hydroxyphenylacetic Acid 0.19 0.11 0.16 0.13 Adenosine 0.48 1.45 0.36 0.69 D-Fructose

0.08 0.20 0.09 0.15 D-Lactic Acid Methyl Ester 0.49 1.33 0.58 0.71 β-Methyl-D-Glucoside 0.33 0.45 0.29 0.16 L-Lactic Acid 0.46 1.29 0.42 0.79 γ-Amino-butyric Acid 0.08 0.15 0.07 0.16 Bromosuccinic acid 0.69 1.57 0.51 0.81 Maltose 0.14 0.41 0.13 0.18 α-Keto-glutaric Acid 0.54 0.67 0.45 0.81 Palatinose 0.13 0.07 0.10 0.18 D-Psicose 0.60 1.64 0.66 0.88 Turanose 0.29 0.64 0.27 0.20 D-Malic Acid 0.47 1.49 0.42 0.88 D-Galactose 0.14 0.37 0.15 0.20 L-Serine 0.59 1.26 0.44 0.91 D-Melibiose 0.17 0.65 0.13 0.23 D-Saccharic Acid 0.53 0.66 0.43 0.91 L-Asparagine 0.08 0.47 0.15 0.23 Succinic Acid Mono-Methyl Ester 0.71 1.35 0.43 0.92 α-Methyl-D-Galactoside 0.17 0.28 0.16 0.24 Adonitol 0.38 0.87 0.35 0.95 L-Alanine 0.12 0.37 0.08 0.24 2-Amino Ethanol 0.40 0.85 0.35 0.99 D-Gluconic Acid 0.08 0.37 0.08 0.26 L-Alanyl-Glycine 0.41 1.12 0.45 1.00 2-Keto-D-Gluconic Acid 0.19 0.25 0.15 0.28 Maltitol 0.48 0.75 0.53 1.02 L-Glutamic Acid 0.58 0.36 0.28 0.30 D-Galacturonic acid 0.60 1.06 0.43 1.08 Gentiobiose 0.51 1.58 0.25 0.31 L-Ornithine 0.81 1.21 0.58 1.08 Maltotriose 0.11 0.58 0.13 0.32 β-Hydroxy-butyric Acid 0.66 0.81 0.53 1.09 Quinic Acid 0.22 0.49 0.24 0.34 Succinamic Acid 0.45 1.26 0.37 1.10 D-Cellobiose 0.37 0.24 0.25 0.34 Arbutin 0.67 1.21 0.53 1.20 N-Acetyl-D-Glucosamine 0.27 0.73 0.31 0.34 Glycerol 0.56 0.63 0.48 1.22 L-Pyroglutamic Acid 0.26 0.97 0.24 0.35 D-Glucosamine 0.47 0.91 0.45 1.43 L-Proline 0.51 0.50 0.46 0.37 α-D-Glucose 0.83 1.49 0.37 1.47 D-Xylose 0.42 0.66 0.29 0.37 L-Malic Acid 0.81 1.54 0.66 1.54 D-Trehalose 0.64 0.65 0.29 0.38 D-Melezitose 0.18 1.22 0.29 0.39 D-Sorbitol

< 0.1 0.1-0.3 0.3-0.5 0.5-0.7 0.7-0.9 0.9-1.1 1.1-1.3 1.3-1.5 1.5-1.7

Figure 4.11: Heatmap of growth of the pathogen FocR1 and the non-pathogenic isolates npM, npC, npD on 95 carbon sources. Values show the average optical density (n = 3). The optical density is ranked from low to high for the pathogenic isolate FocR1.

Interaction of Foc and non-pathogenic F.oxysporum in graminoids

4.4 Discussion

The capacity of Foc to colonize non-susceptible hosts was similar to that of the non- pathogenic F. oxysporum isolates. However, in competition, the non-pathogenic isolates were more effective in colonizing the roots of graminoids in comparison to Foc. Especially after co-inoculation of all four isolates, Foc could barely be traced back in the roots. No difference in root colonization was found between the two plant species, while only the shoot of C. iria was substantially colonized. Root colonization by non-pathogenic isolates is described to be limited to the outer cortex of the roots (Olivain et al., 2003), which might be a non-specific area to colonize. However, root colonization of symptomless hosts by F. oxysporum f. sp. cepae and F. oxysporum f. sp. melonis has been shown to be plant species-dependent (Gordon et al., 1989; Leoni et al., 2013). The shoot colonization of C. iria suggests that in this plant species F. oxysporum can colonize beyond the root cortex and migrate to upper parts (Gordon and Martyn, 1997). Despite the difference in colonization of the areal parts between the plants, this did not result in differential root colonization in this case.

The non-pathogenic F. oxysporum populations in soil were both lower (npC) and higher or equal (npM, npD) than the pathogenic population, as observed by plating. Therefore, differences in root colonization can be largely attributed to the capacity to colonize the roots, but not the soil. The non-pathogenic and pathogen populations in soil when monitored by qPCR were generally not different from each other, but the qPCR results differed from the results obtained with plating. Propagule counts after plating only show the active propagules in soil, whereas DNA based techniques accounts for all active, dormant and dead propagules. It has been reported that free DNA can persist in soil for up to two years in certain conditions (Levy-Booth et al., 2007).

Both the rhizosphere and rhizoplane are environments which are richer in nutrients than the bulk soil. Plants release 5-21% of their photosynthetically fixed carbon as root exudates which contain soluble sugars, amino acids or secondary metabolites (Huang et al., 2014). The competitive ability to colonize roots can be partly determined by the efficiency to retrieve host carbon (Garrett, 1970). Interestingly, the average growth of the pathogenic isolate FocR1 on 95 different carbon sources was not lower, and limited growth was not observed on more carbon sources compared to the non-pathogens. Other aspects probably may have caused the lower competitiveness of the pathogen, such as low tolerance to fungistatic substances excreted by other microorganisms, low tolerance to antimicrobial compounds inside plant tissues and low speed of germination of the chlamydospores (Garrett, 1970;

Chapter 4

Hardoim et al., 2008; Kennedy et al., 2011). Garett (1970) proposed that the continued evolution of the host-parasite relationship reduces competitive saprophytic ability of vascular pathogens.

Trehalose, a storage compound in fungal cells, was the sole carbon source on which all isolates showed very good growth. The other carbon sources that sustained good growth differed among isolates. This could be an indication of habitat adaptation. The pathogenic isolate FocR1 showed especially good growth on L-proline, L-pyroglutamic acid and D- cellobiose in comparison to the other isolates. Proline accumulates in plants suffering drought stress (Qamar et al., 2015), and Foc induces drought stress in the banana plant. Increased growth on proline and pyroglutamic acid, a metabolite related with the proline pathway, could have developed after a prolonged intimate host-pathogen relation. To investigate if this characteristic is present among more Foc isolates, growth on amino acids was compared in a larger collection of Foc and non-pathogens. Profiles differed largely between isolates, but no distinction could be made based on pathogenicity (Chapter 3 and unpublished results). Steinberg et al. (1999) tested F. oxyporum f. sp. lycopersici, F. oxysporum f. sp. radicis-lycopersici and non-pathogenic isolates and concluded that carbohydrate utilization was unique to each isolate and not dependent on pathogenicity.

As Foc isolates are not monophyletic, it is assumed that pathogenicity on banana has evolved several times independently (O’Donnell et al., 1998). Therefore, metabolic adaptations of one isolate can differ from others. In addition, it has been suggested that non- pathogenic isolates can attain pathogenicity through horizontal gene transfer (Ma et al., 2010; Chapter 3), resulting in Foc isolates that have not gone through an extensive period of coevolution in close relation with banana. The Foc isolate used in this study had an identical EF-1α and IGS sequences as Foc isolates from Honduras, Australia, Tanzania, Malaysia, USA and Ghana (Fusarium ID). The worldwide occurrence of the same genotype suggests that this isolate possesses a long term pathogenicity and close relation with banana. The growth profile of Foc isolates with the same EF-1α/IGS sequence type were tested and were highly similar (Chapter 3). It needs to be tested if other Foc isolates, especially those that acquired pathogenicity recently, show the same reduced competitiveness to colonize symptomless hosts as the isolate used in this study.

Interaction of Foc and non-pathogenic F.oxysporum in graminoids

Remarkably, npD, the isolate with the lowest growth on the carbon sources, showed the best root colonization. Its root colonization was not reduced in combined inoculation compared to single inoculation. On the contrary, the root colonization by the isolates that had the highest growth on the carbon sources, namely isolate npC and FocR1, was the most reduced in combined inoculation. We hypothesize that isolate npD, and to a lesser extent npM, are isolates with better endophytic properties compared to npC and FocR1. Isolate npD, together with npM, was encountered abundantly in the shoot of C. iria in both experiments, while FocR1 was only detected in the shoot in one of the two experiments and npC was detected in only one out of eight samples. It has been observed that endophytic bacteria grow better on nutrient poor media, suggesting that an endophytic population has been adapted to low nutrient conditions (Mano et al., 2007). Low growth on abundant carbon sources can be an indication of better growth on low nutrient content. Demers et al. (2015) suggested a lack of specialization of endophytic F. oxysporum isolates, since they did not observe a distinct endophytic and soil F. oxysporum population based on EF-1α sequences. Based on IGS profiles, Edel et al. (1997) recorded that the F. oxysporum population in soil differed from the population associated with tomato and wheat roots, but not with melon and flax. Functional traits of F. oxyporum isolates which have better endophytic properties can be overlooked when evaluating the population based on a single housekeeping gene (Saunders and Kohn, 2009; Chapter 3). We hypothesize that the isolates npD and npM are more adapted to live as endophytes, while isolate npC is rather a saprophyte, having a general higher growth profile. The pathogenic isolate FocR1 showed good growth on few carbon sources, which could be an adaption to the specific banana host.

We have opted to inoculate soil with a chlamydospore rich inoculum to approach natural conditions. In addition, results from preliminary tests with root dipping differed from the soil based experiments (data not shown) and in literature it has been described how root colonization by F. oxysporum differs between hydroponic and soil based systems (Nahalkova et al., 2008; Olivain et al., 2006). Quantification by qPCR was utilized since distinction on colony morphology was occasionally challenging due the presence of other fungi. No 100% accuracy was obtained while analyzing the primers on inoculated root material, but false detection remained small and is very unlikely to have affected the final result. Endophytes can be defined as the microorganisms that can be isolated from surface-disinfected plant tissue or extracted from within a plant, and that do not visibly harm the plant (Hallmann et al., 1997; Petrini, 1991). The samples analyzed with qPCR were not first surface sterilized, as this seemed to interfere with the DNA extraction. Therefore, the detected levels of F. oxysporum cannot exclusively be attributed to endophytic colonization, but also occupation of

Chapter 4 the rhizoplane. However, both endophytic and epiphytic associations can be of importance to estimate the effect of symptomless hosts on Foc inoculum. A subsample of all samples was plated after surface sterilization, with results supporting the data obtained by qPCR. Therefore, we believe that the majority of F. oxysporum detected by qPCR concerns a genuine endophytic association.

The objective of this study was to increase the understanding of the endophytic association of Foc and symptomless hosts. Better understanding is necessary to consider the impact of symptomless hosts on the disease. Our results suggest that the interactions between F. oxysporum isolates can play a greater role in outcome of root colonization on symptomless hosts than the interaction of a single isolate with the plant. The root colonization by the pathogen dropped considerably when non-pathogenic F. oxysporum strains were present. This can be an important aspect to account for in models of population dynamics that aim to estimate pathogen populations. In this study, we considered the interaction between isolates within the F. oxysporum species complex. However, interaction with other plant-associated microorganisms, such as mycorrhiza may also have a considerable impact (Omacini et al., 2006; Pan et al., 2008; Wehner et al., 2011). It is difficult to generalize the observed lower competitive ability of Foc due to the diversity within Foc. If after broader testing of Foc isolates the same pattern of lower of competitiveness of pathogenic isolates can be found, cover crops and weeds can be considered reservoirs of non-pathogenic F. oxysporum, rather than reservoirs of Foc. Since some non-pathogenic F. oxysporum isolates have biocontrol capacities, an increase of those can lower the inoculum potential of Foc (Alabouvette et al., 2009). Therefore, the presence of symptomless carriers will not necessarily worsen the disease. The positive contribution on non-pathogenic F. oxysporum can outbalance the effect as carrier of the pathogen. Pattison et al. (2014) observed that the use of a ground cover of pinto peanut together with the natural vegetation reduced Fusarium wilt of banana. The authors attributed this result to the altered soil biology. Considering the ample means of Foc to persist in soil and the potential influence of other microorganism on plant colonization by Foc, it is in our opinion not relevant to evaluate the effect of symptomless hosts solely based on the capacity to host Foc. A positive or negative contribution of symptomless hosts would be preferably evaluated with on-field trials by monitoring the disease and pathogen population shifts in time.

100

Interaction of Foc and non-pathogenic F.oxysporum in graminoids

4.5 Conclusion

The isolate FocR1 does not have a lower capacity to colonize symptomless hosts, but a lower competitiveness compared to non-pathogenic F. oxysporum isolates. The reduced competitiveness of the pathogen was not due to lower growth on carbon sources. Our results suggest that potential symptomless carriers of the pathogen may rather act as reservoir of non-pathogenic F. oxysporum. Further research is necessary to elucidate these complex interactions.

4.6 Acknowledgements

The authors would like to thank Chien-Jui Huang, Silke Deketelaere and Evelien De Waele for their support to establish the qPCR method and Lisa Heyman to assist with the plants. The study was funded by a PhD scholarship to P.D. from the Fund for Scientific Research- Flanders (FWO).

4.7 Supplementary material

Table S4.1: The mean F. oxysproum isolate density in soil after growth of C. iria or B. decumbens (n = 4). Both the single inoculation with pathogenic isolate FocR1 and combined inoculation of pathogenic isolate FocR1 and a non-pathogenic isolate npM, npC, npD are given. The density of each isolate is given in log(pg g-1 soil). The standard deviation is indicated between brackets. Significant differences between the levels of different F. oxysporum isolates in the same mixture are indicated with an asterisk (paired t-test).

Treatment: C. iria B. decumbens

FocR1 + FocR1 npM npC npD FocR1 npM npC npD

- 3.02 (0.38) 2.47 (0.40) npM 2.77 (0.30) 2.86 (0.19) 2.71 (0.22) 2.50 (0.41) npC 2.53 (0.13) 2.50 (0.13) 2.27 (0.39) 2.25 (0.31) npD 2.26 (0.23) 2.38 (0.34) 2.55 (0.06)* 2.83 (0.13)* All non- 2.27 (0.47) 2.29 (0.35) 2.17 (0.35) 2.48 (0.19) 2.53 (0.04) 2.40 (0.11) 2.28 (0.10) 2.54 (0.10) pathogens

101

Chapter 5

Interaction of banana cultivars with Fusarium oxysporum f. sp cubense and non-pathogenic F. oxysporum

Authors Pauline Deltoura, Soraya C. Françaa, Łukasz Paweł Tarkowskib, Lisa Heymana, Wim Van den Endeb, Monica Höftea

Affiliations aLaboratory of Phytopathology, Department of Crop Protection, Faculty of Bioscience Engineering, Ghent University, Coupure Links 653, 9000, Ghent, Belgium bLaboratory of Molecular Biotechnology of Plants and Micro-organisms, Department of Biology, KU Leuven, Kasteelpark Arenberg 31, 3001 Leuven, Belgium

103

Chapter 5

Abstract

Fusarium oxysporum f. sp. cubense (Foc) is a devastating soil-borne pathogen of banana. It has been proposed that the use of banana cultivar mixtures could moderate losses of Foc, by curtailing inoculum pressure or fostering beneficial microorganisms, such as non- pathogenic Fusarium oxysporum (np-Fox). We studied the interaction of four banana cultivars with Foc race 1 (FocR1) and np-Fox isolates. Cultivar (cv.) Silk (AAB) is highly susceptible to FocR1, cv. Prata (AAB) is moderately susceptible, and cv. Colatina Ouro (AA) and cv. Dwarf Cavendish (AAA) are both resistant. Tissue culture plantlets were grown in autoclaved soil inoculated with (1) the pathogen, (2) a combination of three np-Fox isolates or (3) the pathogen combined with the np-Fox isolates. Root colonization and the soil population were monitored. Only cv. Silk provoked an increase of FocR1 in soil and had higher root colonization by Foc. Cv. Prata, although showing internal symptoms similar to cv. Silk, did not increase FocR1 in soil and had lower root colonization by FocR1. Cv. Colatina Ouro was the only cultivar that caused a consistent increase of np-Fox in soil, while root colonization by np-Fox did not differ between the cultivars. Root exudates collected from the cultivars before or after inoculation with Foc did not affect the germination rate of microconidia of Foc. However, different shifts in the sugar and amino acid content of the root exudates after inoculation with Foc were found. The results suggest that a differential defense reaction of the cultivars is reflected on the composition of root exudates. We speculate that inoculum pressure in soil may be reduced by the use of a mixture of cultivars. Besides the different defense reactions to Foc, other traits, such as the ability to stimulate np-Fox in soil by cv. Colatina Ouro, may be useful to manage the disease. Further field research is needed to reveal the impact of the use of banana cultivar mixtures on Foc inoculum pressure and np-Fox stimulation.

104

Interaction of banana cultivars with Foc and non-pathogenic F. oxysporum

5.1 Introduction

Fusarium oxysporum f. sp. cubense (Foc) is the causal agent of Fusarium wilt, one of the most devastating banana diseases. Foc is subdivided in races 1 to 4, depending on the range of cultivars which it can affect. Control of the disease is challenging since the soil- borne character of the pathogen limits the use of fungicides and biological control has so far been ineffective in the field (Ploetz, 2015). Banana production on infested fields is mostly managed by a shift to resistant cultivars. Historically, problems with Foc race 1 in export plantations were solved by a shift from the use of the susceptible ‘Gros Michel’ to the use of resistant Cavendish cultivars. However, other cultivars, such as ‘Maçã’ (Silk) in Brazil, stay locally popular despite their susceptibility to Foc race 1 due to consumer preference and higher farmers’ profits (Borges and da Silva Souza, 2004; Ploetz, 1990).

Little is known about the genetic background of resistance to Foc in banana. Many defense related genes and signaling pathways in banana are different from model plants such as rice or Arabidopsis (Li et al., 2012; Swarupa et al., 2014). Compatible and incompatible interactions of Foc with the banana plant differ in speed and abundance of the defense responses (Li et al., 2013; Swarupa et al., 2014). In resistant hosts, pathogen attack is effectively stopped by defense reactions such as the production of fungal cell wall degrading enzymes and other pathogenesis related proteins, the production of reactive oxygen species and scavenging, cell wall strengthening via accumulation of phenolic acids, and blockage of the vessels by tyloses and gels (Swarupa et al., 2014).

Many studies have emphasized the importance of F. oxysporum populations in soil on disease development. Firstly, serious losses due to Fusarium wilt are only observed when the pathogen inoculum density in soil has passed a critical threshold (Ben-Yephet et al., 1994; Scott et al., 2014; Stover, 1972). Cavendish cultivars are normally resistant to Foc race 1, but can be affected at high inoculum levels and conducive conditions (Aguilar et al., 2000; Furtado et al., 2009; Stover, 1972). Secondly, the F. oxysporum species complex contains many isolates which do not cause disease, but may be active in biocontrol. Non- pathogenic F. oxysporum isolates can reduce Fusarium wilt by competition for nutrients and infection places and by inducing resistance. The mode of action appears to be strain specific (Alabouvette et al., 2009). The Fusarium wilt suppressive capacity of some soils is attributed to the presence of a large population of non-pathogenic F. oxysporum (Alabouvette et al., 2009).

105

Chapter 5

Plants are one of the major influencers of the microbial population in soil (Garbeva et al., 2004, Kinkel et al., 2011). Some studies have shown the importance of plant genotype on the population of F. oxysporum. Growth of Pueraria javanica as cover crop in oil palm plantations induced an increase in the amount of non-pathogenic F. oxysporum in soil, leading to better disease suppression (Abadie et al., 1997). Specific selection of F. oxysporum isolates by tomato and wheat plants was observed by Edel et al. (1997) and a recent study has shown that different tomato cultivars were associated with a different F. oxysporum population (Demers et al., 2015).

Plants influence the chemical, physical and biological properties of the rhizosphere by exudation of volatile and soluble components, known as root exudates (Bertin et al., 2003). These excretions can contain ions, free oxygen, water, enzymes, mucilage and a diverse array of carbon-containing primary and secondary metabolites (Bais et al., 2006). Root exudation can be both an active and passive process. Active secretion involves the excretion of compounds with known functions, such as lubrication and defense, whereas passive excretion is the gradient-dependent diffusion of metabolites (Bais et al et al., 2006). Up to 40% of the fixed carbon can be excreted, but the amount and composition of root exudates released in the environment is dependent on the plant species, soil, plant age, physiological state, stressors and nutrient availability (Bais et al., 2006; Doornbos et al., 2012). In spite of the importance of root exudates on their environment, very little is known about the root exudates of banana (Buxton, 1962; Kurtz, 2010; Li et al., 2011a).

It seems that root exudates play an important role in the development of Fusarium wilt. Enhanced microconidia germination has been observed in root exudates of banana cultivars susceptible to Fusarium wilt in comparison to resistant cultivars (Buxton, 1962; Li et al., 2011a). Often the disease susceptibility correlates positively with the total amount of nutrients in the root exudates (Buxton, 1962; Hao et al., 2010; Li et al., 2009; Steinkellner et al., 2008). Huang et al. (2008) added root exudates of a melon cultivar resistant to Fusarium wilt to a susceptible cultivar and observed a disease reduction. Additionally, it has been reported that F. oxysporum isolates can modulate the root exudates of the colonized plants (Steinkellner et al., 2008). The germination rate of microconidia of a certain F. oxysporum isolate was higher in the tomato root exudate derived from plants that were colonized by this isolate, compared to the exudate of plants colonized by other isolates.

In a field study in Brazil (Chapter 2), we have observed higher disease suppression in patches with a lower density of the susceptible cultivar Maçã (Silk) and higher diversity of cultivars with some level of resistance to Foc. Cultivars present in the more suppressive

106

Interaction of banana cultivars with Foc and non-pathogenic F. oxysporum locations were either resistant (Ouro and Nanica, a synonym for Dwarf Cavendish) or moderately susceptible (Prata). This result suggests that application of cultivar mixtures might contribute to the control of Fusarium wilt. Cultivars with higher level of disease resistance might limit the increase and spread of the pathogen. However, Scott et al. (2014) compared the root colonization of lettuce varieties with different degree of resistance to Fusarium wilt by F. oxysporum f. sp. lactucae and found no solid relation between pathogen colonization and disease resistance. It is plausible that resistant banana cultivars restrict Foc colonization, but data to support this hypothesis are not available. It is under dispute if cultivars with moderate resistance based on quantitative traits contribute to Foc inoculum increases in soil (Dita, 2015).

We investigated whether banana cultivars with different levels of resistance to Foc would interact differently with F. oxysporum communities. Root colonization by Foc race 1 and three different non-pathogenic F. oxysporum isolates was tested on four banana cultivars with different levels of susceptibility to Foc race 1. In addition, the soil population of F. oxysporum was monitored after growth of the cultivars. Moreover, disease symptoms and root colonization were evaluated when non-pathogenic and Foc were co-inoculated. We hypothesized that differences observed in soil and roots could partly be explained by differences in the root exudates. Root exudates of the four cultivars, with and without infection, were collected and compared for their interaction with F. oxysporum isolates by monitoring germination of F. oxysporum microconidia. In addition, the sugar and amino acid content, which is known to make up the two most abundant classes of metabolites present in root exudates (Moe, 2013), was analysed.

5.2 Material and methods

5.2.1 Fungal isolates and inoculum

The isolates used in this study were collected on a farm in Pedra Dourada in Brazil (Table 5.1). One isolate is a Foc race 1 and the other isolates caused no disease on banana cv. Silk (Chapter 3). Throughout the text, the terms pathogenic and non-pathogenic always refer to the virulence on banana cv. Silk.

107

Chapter 5

Table 5.1: F. oxysporum isolates used in this study

Working name Official namea Origin Pathogenicityb FocR1 FoxPDa15 Banana cv. Maçã f. sp. cubense Race 1 npM MrPDb1 Banana cv. Maçã np npC GrPDa1b Cyperus iria np npD GsPDb1c Digitaria insularis np a Name as listed in chapter 3 b np: non-pathogenic on banana cv. silk

A chlamydospore-rich inoculum was prepared according to the method of Smith and Snyder (1971). Autoclaved air-dried soil was inoculated with a conidiospore suspension made from a 10 day old culture grown on Potato Dextrose Agar (PDA). The inoculated soil was allowed to dry for 3 weeks. Final inoculum level, determined by dilution plating on PDA amended with 100 µg l-1 streptomycine, was generally in the range of 105 cfu g-1 soil.

5.2.2 Root and soil colonization of F. oxysporum isolates

Four banana cultivars with different susceptibility to Foc race 1 were compared for their interaction with different F. oxysporum isolates (Table 5.2). Accessions were kindly provided by the International Musa Germplasm Transit Centre (ITC , Leuven, Belgium) and multiplied via tissue culture. Sandy soil collected in Denderbelle, Belgium, was autoclaved for 1 h during 2 consecutive days. Eight week old plantlets of the different cultivars were planted in the sterilized soil which had been inoculated with a chlamydospore-rich inoculum of F. oxysporum. The experiment existed out of three treatments with different inoculum: (1) 5.103 g-1 soil of pathogenic isolate FocR1, (2) a combination of three non-pathogenic isolates, npM, npC, npD, each in concentration of 5.103 g-1 soil (total concentration 15.103 g-1 soil) and (3) a combination of pathogenic isolate FocR1 and three non-pathogenic isolates npM, npC, npD, each in concentration of 5.103 g-1 soil (total concentration 20.103 g-1 soil). As negative control, plantlets were planted in non-inoculated soil. Four plantlets were used per treatment. After inoculation, plantlets were grown for 65 days in a growth chamber (26°C, 16/8 light regime). Plants were weekly evaluated for external disease symptoms: yellowing of the leaves, splitting of the pseudostem and stunted growth. After 65 days, plants were uprooted, and residual soil was carefully removed from the roots. Internal symptoms were assessed by using the Rhizome Discoloration Score (RDS), based on the discolored area of the stellar region (0: no discoloration of the stellar region; 1: 1-25%; 2: 26-50%; 3: 51-75%; 4: 76-100%) (Adapted from Mak et al., 2004). Two centimeter below the rhizome, a piece of 1 cm was cut of all primary roots, which was stored at -20°C for DNA extraction. Additionally, a soil sample was taken for dilution plating on Fusarium selective Komada medium (Komada, 1975). The

108

Interaction of banana cultivars with Foc and non-pathogenic F. oxysporum soil populations from the different cultivars were compared by Kruskal Wallis test (p = 0.05). This experiment has been performed twice. The repetitions are called experiment 1 and experiment 2 throughout the manuscript.

Collected root pieces were crushed in liquid nitrogen and DNA extraction was performed on a 70 – 90 mg sample with the Invisorb spin plant mini kit (Stratec Molecular, Berlin) following the manufacturer’s instructions. Quantification of root colonization was done with qPCR with isolate specific primers using the primers and conditions as described in Chapter 4. Root colonization was compared between the cultivars by a Kruskal Wallis test (p = 0.05).

Table 5.2: Banana cultivars used in this study (Borges et al., 2009)

Cultivar Brazilian Genomic Subgroup Susceptibility to Foc race 1 ITC accession synonym group code Silk Maçã AAB Silk Very susceptible ITC0348 Prata Prata AAB Pome Moderately susceptible ITC0207 Colatina Ouro Colatina Ouro AA Unknown Resistant ITC0260 Dwarf Cavendish Nanica AAA Cavendish Resistant ITC0002

5.2.3 Root exudate collection

Tissue culture plantlets of the cultivars Silk, Prata, Colatina Ouro and Dwarf Cavendish were grown in a growth chamber (26°C, 16/8 light regime). Plants were grown on a mixture of sand and vermiculite (1:1) since this substrate is inert and easily removable. After 106 days, the plantlets were inoculated with Foc race 1 (isolate FocR1) by a 2 h root dip in a spore suspension (5. 105 spores ml-1). Non-inoculated plants were dipped in sterile water. Four plants were used per treatment. After root dipping, the plants were replanted in a sand vermiculite mixture (1:1) and placed in the growth chamber. One week after inoculation, the plants were uprooted and the substrate carefully removed avoiding root injures. The plantlets were placed in glass jars and the roots submerged in sterile distilled water. The pots were wrapped in aluminum foil to create a dark environment for the roots and subsequently placed in the growth room. After 24 h, the root exudates were collected, filtered over a paper filter to remove solid particles and the volume was registered. Immediately after collection, the root exudates were filter sterilized (Millipore, mesh 0.45 µm) and stored at -20 °C. A part of the exudates was kept in small quantities to be defrosted for germination counts. The part for chemical analysis was immediately freeze dried. Fresh root weight, shoot weight, and internal symptoms of the plantlets were determined.

109

Chapter 5

5.2.4 Germination of microconidia in root exudates

Exudates were diluted to a final concentration of 10 ml g-1 root fresh weight. Of each exudate three wells of a 24 well-plate were filled with 500 µl. As control, four times 3 wells were filled with sterile distilled water. The samples were inoculated with 100 µl of a 107 spore ml-1 spore suspension of one of the isolates FocR1, npM, npC, npD, prepared from 14 day old cultures on PDA. The plates were incubated gently shaking at 28 °C for 24 h. A total of 200 microconidia per sample were checked for germination under the light microscope. A spore was considered germinated when the germination tube reached at least the length of the spore. The germination rates in the different exudates were compared with a post-hoc Tukey test.

5.2.5 Chemical composition of the root exudates

Freeze dried exudates were diluted in 200 µl water and heated for 5 min at 90 °C, in order to stop possible enzymatic activities which may have an impact on the chemical composition of the samples. A 50 µl aliquot of this diluted exudate was passed through a two layers-bed ion- exchange column composed of 100 µl bed volume of Dowex-50® H+ and a 100 µl bed volume of Dowex®-1-Acetate to obtain the neutral fraction containing soluble carbohydrates. The resins were rinsed six times with 50 µl MilliQ water. A 50 µl of this sample was supplemented with 50 µl mannitol solution (20 µM) with 0.02% azide. Mannitol was here used as internal standard for chromatographic data analysis. Carbohydrates were separated by anion exchange chromatography and quantified by integrated pulsed amperometric detection (HPAEC-IPAD, Thermo Fisher Scientific, Waltham, Massachusetts, USA). A CarboPacTM PA100-column (2mm x 50mm) and an analytic CarboPacTM PA100-column (2mm x 250mm) were used in series circuit. The column is equilibrated with 90 mM NaOH before injection. The sugars were eluted with a Na-acetate gradient: 0 mM to 10 mM from 0 min to 6 min; 10 mM to 100 mM from 6 min to 16 min. The flow rate is 0.25 ml min-1. Finally, the column was regenerated with 500 mM Na-acetate for 5 min. Elution time of each sugar was determined by injecting pure compounds. Quantification of the sugars was performed with the use of external standards injected at a known concentration of 10 µM. Data were processed with the Chromeleon™ software, version 6.70, from DionexTM.

For amino acid analysis, 10 mg of resin C18 was added to 50 µl of the heated exudate solution. The samples were vortexed and agitated at 1050 rpm during 1 h. After centrifugation, the supernatant was recovered and diluted 6 times with a mannitol solution (20 µM) with 0.02% azide. Nor-valine was added as internal standard (25 µM). Amino acids

110

Interaction of banana cultivars with Foc and non-pathogenic F. oxysporum were derivatized as O-PhtalAldehydes (Altmann, 1992) and subsequently detected in a Shimadzu Prominence 20A HPLC system with Shimadzu fluorescence detector (RF-10Axl) and automatic injector with pretreatment function (Smartline Autosampler 3950). Derivatized amino acids were separated on a YMC Triart C18 column (ID3.1 x 250 mm, 3 µm particle size) with the use of two buffers: a 50 mM phosphate buffer pH=6.5 containing 0.7% (v/v) tetrahydrofuran, referred as solvent A; a solution made with 45% acetonitrile, 40% MeOH and 15% H2O, referred as solvent B. Derivatized amino acids were eluted as follows: solvent A gradient from 96% to 45% from 0 to 50 min, 100% solvent B from 50 to 55 min. The column was regenerated by washing it in 96% solvent A for 5 min. Elution time of each amino acid was determined by injecting pure compounds. Data were processed in Clarity Chrom software, version 2.6.2.212 (Knauer). Quantification of amino acids was performed with the use of internal standards injected at a known concentration of 25 µM.

The amounts of sugars and amino acids with and without inoculation within one cultivar were compared with a Wilcoxon rank-sum test.

5.3 Results

5.3.1 Interaction between Foc and banana cultivars

After inoculation with Foc race 1 isolate FocR1, cultivars Silk and Prata showed similar discoloration of the rhizome (Figure 5.1A). However, root colonization by the pathogen was significantly higher in cv. Silk than in cv. Prata. No discoloration of the rhizome was observed in the resistant cultivars Colatina Ouro and Dwarf Cavendish (Figure 5.1A), but the pathogen could be found back in low amounts in the roots of both cultivars (Figure 5.1B). Only cv. Silk seemed to stimulate propagules of FocR1 in soil (Figure 5.1C).

111

Chapter 5

A Experiment 1 Experiment 2

4 4

3 4 3 4 3 3 2 2 2 2

1 1 1 1 Number of Plants of Number 0 0 0 0 Silk Prata OuroCol O NanicaD Cav Silk Prata OuroCol O Nanica D Cav

B 6 6

a 5 b 5 b b 4 4 b b b

3 3

1 root fresh weight) fresh root 1

- - 2 2

1 1

Log (pg DNA g DNA (pg Log Log (pg DNA g DNA (pg Log 0 0 Silk Prata Col O D Cav Silk Prata Col O D Cav

C 1800 a 900 a 1600 800

1400 700

1200 600

soil

1 - - 1000 500

800 400 Cfu Focg Cfu Cfu Focg Cfu 600 b 300 b b a 400 200 a 200 100 a 0 0 Silk Prata Col O D Cav Silk Prata Col O D Cav

Figure 5.1: Banana cultivars inoculated with the pathogenic isolate FocR1 (n = 4). A: internal disease symtoms: the Rhizome Discoloration Score (0: no discoloration of the stellar region; 1: 1-25%; 2: 26-50%; 3: 51- 75%; 4: 76-100% ). B: Root colonization by the pathogen. Each diamond shows one data point. Grey diamonds show values below the analytical detection limit. C: Foc propagules retrieved from soil after growth of a cultivar. Bars show the median and error bars show the MAD (median absolute deviation). Different letter show significant differences (Kruskal Wallis test, p = 0.05). Silk (AAB), Prata (AAB), Col O: Colatina Ouro (AA), D Cav: Dwarf Cavendish (AAA).

112

Interaction of banana cultivars with Foc and non-pathogenic F. oxysporum

5.3.2 Interaction of banana cultivars with non-pathogenic F. oxysporum isolates

In the first experiment, root colonization of cultivar Colatina Ouro by the non-pathogenic F. oxysporum isolates was slightly lower compared to the other cultivars. However, in the second experiment, root colonization by non-pathogenic F. oxysporum isolates was not different among cultivars (Table 5.3). There was no consistent difference between isolates and all isolates were detected above the detection limit in all root samples. The F. oxysporum soil population was larger in soil in which cv. Colatina Ouro had grown compared to the other cultivars (Figure 5.2). In soil, the proportions of the different isolates did not differ significantly (data not shown).

Table 5.3: Root colonization (log (pg DNA g-1 root fresh weight)) of banana cultivars by non-pathogenic F.oxysporum isolates (npM, npC, npD). The standard deviation is given between brackets. Different letters show significant differences within rows, within each experiment (Kruskal Wallis, p = 0.05, n = 4). Silk (AAB), Prata (AAB), Colatina Ouro (AA), Dwarf Cavendish (AAA).

Experiment 1 Experiment 2 Colatina Dwarf Colatina Dwarf Silk Prata Silk Prata Ouro Cavendish Ouro Cavendish npM 4.07 3.95 3.42 4.35 3.88 3.86 3.59 3.65 (0.21) a (0.43) a (0.34) b (0.16) a (0.16) a (0.55) a (0.26) a (0.21) a npC 3.90 3.84 3.41 4.25 3.78 4.05 3.76 3.80 (0.20) ab (0.42) bc (0.25) c (0.22) a (0.10) a (0.31) a (0.10) a (0.18) a npD 4.48 4.41 3.97 4.71 3.90 4.17 3.95 3.76 (0.19) a (0.35) a (0.08) b (0.37) a (0.13) a (0.46) a (0.18) a (0.27) a

4000 Experiment 1 a 16000 Experiment 2 a 3500 14000

3000 12000

b ab

2500 10000

Soil

1 1

1 - - b 2000 8000

1500 6000 b

Cfu Fox g Fox Cfu Cfu Fox g Fox Cfu 1000 4000 c 500 2000 c 0 0 Silk Prata Col O D Cav Silk Prata Col O D Cav

Figure 5.2: F. oxysporum propagules retrieved from soil after growth of a cultivar in soil inoculated with non- pathogenic F.oxysporum isolates. Bars show the median cfu of the sum of all F. oxysporum isolates and error bars show the MAD (median absolute deviation). Different letter show significant differences (Kruskal Wallis test, p = 0.05, n = 4). Silk (AAB), Prata (AAB), Col O: Colatina Ouro (AA), D Cav: Dwarf Cavendish (AAA).

113

Chapter 5

5.3.3 Influence of non-pathogenic F.oxysporum on the interaction Foc-banana cultivars

Root colonization of the banana cultivars by the pathogenic isolate FocR1 was monitored after single inoculation and combined inoculation with the non-pathogenic F. oxysporum isolates (Table 5.4). In the first experiment, the presence of non-pathogenic F. oxysporum significantly decreased Foc colonization in the roots of cultivar Silk. However, in the second experiment, the reduction was not significant due to large variation. Also in other cultivars a slight decrease in root colonization by pathogenic isolate FocR1 was observed after co- inoculation with non-pathogenic isolates, although not significant. A figure showing the different data points can be found in the supplementary material S5.1.

The presence of non-pathogenic isolates delayed the disease in cultivar Prata. The disease progression in cultivar Silk did not differ between single inoculation and co-inoculation with non-pathogenic F. oxysporum (Figure 5.3). Due to the limited number of plants, no strong statistical support could be obtained.

Table 5.4: Root colonization by the pathogenic isolate FocR1 (Foc) when single inoculated and co-inoculated with the three non-pathogenic isolates (np-Fox) (log (pg DNA g-1 root fresh weight). An asterisk indicates significant difference between Foc and Foc+np-Fox for each cultivar (Wilcoxon rank-sum test, p = 0.05, n = 4). Silk (AAB), Prata (AAB), Colatina Ouro (AA), Dwarf Cavendish (AAA).

Silk Prata Colatina Dwarf Ouro Cavendish Experiment 1 Foc 4.97 (0.62) 3.43 (0.97) 2.94 (0.26) 3.34 (0.35) Foc +np-Fox 3.44 (0.30) * 2.95 (0.79) 2.69 (0.24) 3.29 (0.31) Experiment 2 Foc 4.66 (0.36) 3.25 (0.57) 2.79 (0.36) 2.96 (0.12) Foc +np-Fox 3.84 (1.12) 3.21 (0.25) 2.43 (0.72) 2.16 (0.54) *

114

Interaction of banana cultivars with Foc and non-pathogenic F. oxysporum

Foc Foc + np-Fox

4 4

3 3 Silk 2 2

1 1

Diseased plants Diseased 0 0 0 6 13 20 27 34 42 47 53 58 65 0 6 13 20 27 34 42 47 53 58 65

3 3 Experiment 1 Experiment

2 2 Prata Prata 1 1

Diseased plants Diseased 0 0 0 6 13 20 27 34 42 47 53 58 65 0 6 13 20 27 34 42 47 53 58 65

4 4

3 3

2 2

Silk

1 1 Diseased plants Diseased 0 0 0 7 14 21 28 35 42 49 56 65 0 7 14 21 28 35 42 49 56 65

Experiment 2 Experiment 3

2 2 Prata Prata 1 1

Diseased plants Diseased 0 0 0 7 14 21 28 35 42 49 56 65 0 7 14 21 28 35 42 49 56 65

Dpi Dpi Figure 5.3: Disease progression curves of cv. Silk (AAB, subgroup Silk) and cv. Prata (AAB, subgroup Pome) when inoculated with Foc alone and with Foc combined with non-pathogenic F.oxysporum (np-Fox).

5.3.4 Root exudates

Root exudates of the cultivars, with and without Foc inoculation, were collected and their sugar and amino acid content was analysed. The total amount of sugar in the root exudates was not statistically different among the non-inoculated cultivars. Only minor differences in the composition were found. Arabinose levels in the exudate of cultivar Silk were significantly higher than those of cv. Dwarf Cavendish and the xylose content was higher in the exudate of Colatina Ouro than in the exudates of cvs. Prata and Dwarf Cavendish (statistics on the comparison of non-inoculated cultivars can be found in supplementary material S5.2).

115

Chapter 5

After Foc inoculation, the total sugar content in the root exudates significantly increased in the cultivars Silk, Colatina Ouro and Dwarf Cavendish (Figure 5.4A). Generally, the level of all sugars increased after inoculation, although to different extent. In cultivar Silk, the sucrose concentration in the exudate after Foc inoculation was 6 times higher than without inoculation. Also the level of rhamnose and glucose almost doubled after inoculation. In the exudate of cultivar Prata, only small, non-significant increases were observed. In the exudate of cultivar Colatina Ouro, the levels of rhamnose and arabinose tripled, although with high variation. Also mannose, glucose and sucrose doubled after inoculation. The highest increase in sugar content after inoculation was observed in the exudate of the resistant cultivar Dwarf Cavendish. The average level of arabinose increased ninefold, while the average of other sugars doubled (Figure 5.4B).

In contrast to the total sugar content in the exudates, the total amount of amino acids was different between the non-inoculated cultivars. Cv. Colatina Ouro excreted significantly less amino acids in comparison to the other cultivars (Figure 5.5). After Foc inoculation, only the amino acid content in the exudate of cultivar Colatina Ouro increased significantly and was no longer different from the other cultivars (Figure 5.5). Only small differences in the composition of the different amino acids were found among the different cultivars (Figure 5.6). In cultivar Silk, glutamic acid levels in the exudate decreased significantly after inoculation. In cultivar Prata, glutamic acid levels in the exudate increased, while phenylalanine levels decreased after Foc inoculation. It is remarkable that GABA was almost absent in the exudate of the non-inoculated plants of cv. Colatina Ouro, while after Foc inoculation GABA was the second most abundant amino acid, after alanine. In exudate of cultivar Dwarf Cavendish, glutamic acid and proline increased after inoculation, but large variations were observed. When comparing the proportions of each amino acid in the exudates, only small differences were found between the cultivars (Supplementary material S5.3).

The germination of Foc microconidia was not different in the exudate of the cultivars, or in the exudate of plants inoculated with Foc. A different germination rate in the exudates was only observed for non-pathogenic isolate npC. Generally, the germination of the microconidia was higher in water than in the root exudates (Table 5.5).

116

Interaction of banana cultivars with Foc and non-pathogenic F. oxysporum

900 A * 800 * * 700

600 root

500 1 - Foc - 400

µmol g µmol Foc + 300 200 100 0 Silk Prata Colatina Dwarf Ouro Cavendish

B 300 Silk 300 Colatina Ouro 250 250

200 * 200

root root

1 1 -

- 150 * 150 µmol g µmol µmol g µmol 100 100 * 50 50

0 0

300 Prata 300 Dwarf Cavendish *

250 250

200 200

* root

root 300

250

1 200 1 - 150

- 150 150 100500

root

1 100 100 - µmol g µmol * µmol g µmol * * 50 50

0 0 µmol g

Foc - Foc +

Figure 5.4: Total sugars and sugar composition of root exudates in µmol g-1 root of four banana cultivars (Silk (AAB), Prata (AAB), Colatina Ouro (AA) and Dwarf Cavendish (AAA)) when not inoculated, and one week after inoculation with Foc race 1 (Foc - and Foc + respectively). A: Total sugar content in the root exudates. B: Composition of the sugar content. Significant differences between the non- inoculated and inoculated treatment are indicated with an asterisk (p = 0.05; Wilcoxon rank-sum test, n = 4). Error bars show the standard deviation.

117

Chapter 5

9 * 8 7

root

1 5 - Foc - 4 Foc + mmol g mmol 3 2 1 0 Silk Prata Colatina Ouro Dwarf Cavendish

Figure 5.5: Total amount of amino acids in root exudates of four banana cultivars, non-inoculated and inoculated with Foc (Foc – and Foc +, respectively). Significant differences between the non-inoculated and inoculated treatment are indicated with an asterisk (p = 0.05; Wilcoxon rank-sum test, n = 4). Silk (AAB), Prata (AAB), Colatina Ouro (AA), Dwarf Cavendish (AAA).

Figure 5.6: Amino acid composition of root exudates in µmol g-1 root of four non-inoculated banana cultivars (Silk (AAB), Prata (AAB), Colatina Ouro (AA) and Dwarf Cavendish (AAA)), and one week after inoculation with Foc race 1 (Foc - and Foc + respectively). Significant differences between the non-inoculated and inoculated treatments are indicated with an asterisk (p = 0.05; Wilcoxon rank-sum test, n = 4). On cultivar Colatina Ouro, no statistics have been performed on the absolute values, due to the large difference in total amino acid exudation. Error bars show the standard deviation.

118

Interaction of banana cultivars with Foc and non-pathogenic F. oxysporum

1200 Silk

1000

root 800

1 - 600 *

µmol g µmol 400 200 * 0

1200 Prata 1000 *

root 800

- * 600

µmol g µmol 400 200 0

Colatina Ouro

150 Foc -

root 1 1 - 100 1800 Colatina Ouro 50 1600 g µmol 1400

lys

tyr

his

gly

val

ala

thr

glu gln

ser

leu

arg

asp asn

orn pro

- phe

1200 met

GABA

iso-leu

root

1 1000 - 800

µmol g µmol 600 400 200 0

1200 Dwarf Cavendish

1000

800

root

1 - 600

µmol g µmol 400

200 * * * 0

Foc - Foc +

119

Chapter 5

Table 5.5: Average rate germination of microconidia (%) after 24h in root exudates of four different banana cultivars. Exudates have been collected from non-inoculated plants, or one week after inoculation with Foc race 1 (Foc – and Foc +, respectively) and have been diluted to 1 ml per 10 g roots. The standard deviation is shown between brackets. Different letters indicate significant differences in germination of a single isolate based on a post-hoc Tukey test (n = 4, p = 0.05). Silk (AAB), Prata (AAB), Colatina Ouro (AA), Dwarf Cavendish (AAA).

Germination rate Cultivar FocR1 npM npC npD

water 8.93 (0.58) a 23.89 (4.26) a 8.92 (3.88) abc 44.91 (9.05) a

Silk Foc - 5.15 (1.75) b 14.43 (2.56) b 10.74 (2.70) ab 14.67 (5.57) b Foc + 3.68 (0.53) b 12.38 (1.97) b 4.01 (1.11) cd 14.48 (2.80) b

Prata Foc - 5.69 (1.20) b 14.24 (1.96) b 13.40 (5.49) a 14.68 (4.85) b Foc + 3.87 (0.89) b 10.44 (1.23) b 2.38 (0.76) d 13.41 (4.75) b

Colatina Ouro Foc - 3.84 (0.67) b 10.69 (1.04) b 1.83 (0.85) d 16.74 (6.20) b Foc + 5.06 (1.40) b 14.78 (2.60) b 5.72 (1.22) bcd 12.93 (2.61) b

Dwarf Foc - 5.09 (0.97) b 13.30 (1.38) b 2.88 (1.21) cd 14.51 (3.64) b Cavendish Foc + 4.78 (0.95) b 14.91 (1.65) b 5.52 (1.41) bcd 13.28 (1.55) b

Table 5.6: Overview of the characteristics of the different cultivars.

Silk Prata Colatina Ouro Dwarf Cavendish

F. oxysporum Disease symptoms Yes Yes No No Foc in roots High Low Low Low Foc stimulation in soil Yes No No No Np-Fox in roots Yes Yes Yes Yes Np-Fox stimulation in soil No No Yes No Np-Fox suppresses Foc root ± No No ± colonization Disease suppressed by np-Fox No Yes

Root exudate

Change total sugars after ↗ = ↗ ↗ inoculation Average increases of more than sucrose (6.3 x) rhamnose (3.53 x) arabinose (9.03 x) 50%: glucose (1.8 x) arabinose (3.02 x) sucrose (2.36 x)

rhamnose (1.8 x) mannose (2.56 x) mannose (2.31 x)

glucose (2.32 x) glucose (2.27 x)

sucrose (2.28 x) xylose (2.23 x)

Total AA without Foc high high low high Change total AA after inoculation = = ↗ = Changes in AA composition after Glu ↘ Glu ↗ GABA ↗ Glu ±↗ inoculation Phe ↘ Pro ±↗

120

Interaction of banana cultivars with Foc and non-pathogenic F. oxysporum

5.4 Discussion

In this study, we show that banana cultivars Silk, Prata, Colatina Ouro and Dwarf Cavendish can interact differently with pathogenic and non-pathogenic F. oxysporum isolates. Root colonization by the pathogen corresponded to the susceptibility of the cultivars, but the root colonization by non-pathogenic isolates was equal in all cultivars. One of the cultivars, Colatina Ouro, seemed to stimulate the non-pathogenic soil population. Infection by Foc caused different shifts in the sugar and amino acid composition of the root exudates of the cultivars. Probably these shifts are the outcome of diverse defense reactions. Cultivar differences concerning the interactions with Foc and non-pathogenic F.oxysporum isolates (np-Fox) and root exudates are summarized in Table 5.6.

The level of root colonization by FocR1 correlated positively with the level of susceptibility to Foc race 1 described in literature and confirmed in this study. Resistant cultivars to Foc are not immune; the pathogen can colonize their roots (Li et al., 2013; Pegg and Langdon, 1987). Our study shows that root colonization of the resistant cultivars Colatina Ouro and Dwarf Cavendish is limited. Remarkably, cv. Prata, which is moderately susceptible, showed the same level of rhizome discoloration as the very susceptible cv. Silk, but had significant lower root colonization and did not stimulate the Foc population in soil. Additionally, co-inoculation with np-Fox isolates could delay the disease in cv. Prata, in contrast to cultivar Silk. These results suggest that cv. Prata displays a resistance reaction in the early phase of the infection, limiting root colonization. However, after invasion, cv. Prata seems to be unable to stop the pathogen from colonizing the rhizome and pseudostem. Li et al. (2012) proposed that resistance of banana towards Foc race 4 is expressed before root colonization by the pathogen as they observed less spores attached to roots of resistant cultivars compared to susceptible cultivars. Analysis of microconidia germination in the root exudates showed inhibition in the exudate from resistant cultivars (Buxton, 1962; Li et al., 2011a). Buxton (1962) attributed this difference to higher sugar and amino acid content in the root exudate of the susceptible cultivar compared to the resistant cultivar.

Root exudates are the first contact of the pathogen with the plant. To explore this interface, the sugar and amino acid content of root exudates of non-inoculated and inoculated plants were compared. After inoculation with Foc a significant increase in the sugar level in the exudates of resistant cultivars was observed. It is well described that plant defense reactions to pathogen invasion impact heavily on the carbohydrate metabolism of the plant (Berger et al., 2007; Bolton, 2009; Rojas et al., 2014). The defense reaction is energy demanding and requires extra availability of carbon skeletons for production of secondary metabolites. In

121

Chapter 5 addition, sugars, especially glucose, sucrose and fructose, are known to act as signal molecules to activate defense pathways (Rojas et al., 2014). Infected tissue turns from source to sink, or in the case of roots, becomes a higher demanding sink (Berger et al., 2007; Bolton, 2009). In an incompatible interaction of a pathogen with a resistant host, this reaction tends to be stronger than in a compatible interaction with a susceptible host. The increased sugar level in the root exudates of the resistant cultivars can result from leakage from the upregulated carbohydrate metabolism in the roots due to the defense reaction. Benhamou et al. (1991) observed that the incompatible reaction of F. oxysporum f. sp. lycopersici with tomato resulted in a higher increase of invertase activity in the apoplast compared to the compatible interaction. An increased invertase activity promotes higher sucrose transport to the roots and unloading from the phloem via cleavage into glucose and fructose.

Also in the susceptible cultivar Silk, an increase in sugars in the exudate was observed after inoculation. In a compatible interaction, besides the reaction of the plant, the pathogen can also manipulate the carbohydrate metabolism of the plant, for instance by the sucrolytic enzyme arsenal of the pathogen (Bolton, 2009). In the exudate of cv. Silk mainly the sucrose level increased, while in the cvs. Colatina Ouro and Dwarf Cavendish a higher increase in arabinose was observed. The increase of different sugar compounds can be explained by a higher influence of the pathogen in cv. Silk, while the resistance reaction of the plant in the root exudates of resistant cultivars may have a bigger contribution. One of the main resistance reactions of banana plants to Foc is the blockage of the vessels by tyloses, accompanied by the production of pectin rich gels, callose and gums to completely seal of the vessel (Pegg and Langdon, 1987; Vander Molen et al., 1982; Yadeta and Thomma, 2013). Pectin in banana plants is rich in arabinose (Happi Emaga et al., 2008). An increased pectin production to seal the vessels or strengthen the cell walls may have caused extra leaching of arabinose in the root exudates.

Also small differences in amino acid composition of the root exudates after inoculation, could be caused by different resistance reactions. Glutamate metabolism has a key role in plant defense against pathogens (Seifi et al., 2013). The glutamate content in the exudate of cv. Silk decreased after inoculation with Foc while it increased in the exudates of the cvs. Prata and Dwarf Cavendish. This could be an indication of an effective defense reaction in Prata and Dwarf Cavendish, but not in Silk. The exudate of the non-inoculated cv. Prata had a high level of phenylalanine, which decreased after inoculation with Foc. After infection, phenylalanine from the roots can be less available in the root exudate as it is used by phenylalanine ammonia-lysase (PAL) in the phenyl propanoid pathway for the biosynthesis

122

Interaction of banana cultivars with Foc and non-pathogenic F. oxysporum of polyphenolic compounds, such as lignin for cell wall enforcement (de Ascensao and Dubery, 2000). It would be interesting to further investigate if upregulation of PAL is one of the mechanisms by which cv. Prata achieves higher resistance in the roots, while cvs. Colatina Ouro and Dwarf Cavendish deploy other mechanisms. Also the increase of GABA in the exudate of cv. Colatina Ouro can seem to be an indication of active resistance response, as the GABA shunt is known to be upregulated in response to stresses (Roberts, 2007; Seifi et al., 2013).

Our results suggest that different metabolic adjustments, potentially related to defense processes, such as decrease of phenylalanine or increase in arabinose, can be observed from the analysis of the sugar and amino acid analysis of root exudates. Screening germplasm for resistance by analysis of the root exudates one week after inoculation could be a relatively fast, easy and cheap method. Greenhouse assays based on root dip are often too aggressive to allow reliable evaluation of resistance based on symptoms (Pegg and Langdon, 1987). However, the results need to be confirmed and more cultivars should be tested.

Cv. Colatina Ouro differed from the other cultivars by stimulating non-pathogenic F. oxysporum in soil and by having a very low amino acid content in the root exudates of non- inoculated plants. Some authors describe inhibitory action of amino acids on F. oxysporum (Moe, 2013; Okada and Matsubara, 2012). However, in a BIOLOG essay performed in chapter 4, good growth of F. oxysporum isolates was observed on the most abundant amino acids, namely alanine and GABA. Additionally, germination of microconidia was not stimulated in the root exudate of cv. Colatina Ouro.

The germination rate of F. oxysporum microconidia in the root exudates of the cultivars was similar and can as such not explain the observed differences in the F. oxysporum population in soil. This result is contrasting to the findings of Buxton (1962) and Li et al. (2011a), who found a lower germination rate in the exudate of resistant cultivars compared to susceptible cultivars. In soil however, Foc can be present as chlamydospores, sclerotia or mycelium, which may react differently to root exudates than microconidia (Hage-Ahmet et al., 2013). One non-pathogenic isolate had different germination rates in the exudates, which is an indication of isolate specific interaction of F. oxysporum with banana cultivars. Remarkably, germination of microconidia from most F. oxysporum isolates was higher in water than in the exudates in spite of the presence of sugars and amino acids that can serve as carbon source. We hypothesize that germination is inhibited due to the presence of inhibiting factors in the exudate that outweighs the effect of carbon sources for growth.

123

Chapter 5

In the field study (Chapter 2), we have observed that locations with higher Fusarium wilt suppression had a lower density of cv. Maçã (Silk subgroup) and higher diversity of other cultivars (Prata, Ouro and Nanica). This observation has led to the hypothesis that a mixture of cultivars could contribute to the management of Foc. Our results support this hypothesis, as only the highly susceptible cultivar Silk appeared to contribute to an inoculum increase by higher root colonization and stimulation of Foc in soil. Cultivar Prata, although susceptible to Foc race 1 in the greenhouse often shows field resistance. The low root colonization by Foc and differential root exudates, especially concerning the amino acids, suggest that cv. Prata possesses some quantitative resistance traits at the root level. However, its resistance traits are probably less or less effective than those in resistant cultivars. For example, no shift in the sugar content of the exudate of cv. Prata has been observed after inoculation, in contrast to the resistant cultivars. Non-pathogenic F. oxysporum could delay the disease in cv. Prata, while not in cv. Silk. This suggests that other microorganisms, such as non-pathogenic F. oxysporum, can contribute to the field resistance observed for cultivar Prata, while not, or less in cv. Silk. Besides the influence on the pathogen population, cultivars may possess other interesting traits, such as the capacity to stimulate non-pathogenic F. oxysporum in soil as observed for cv. Colatina Ouro. The analysis of sugars and amino acids already revealed significant differences between the cultivars, suggesting that also other (beneficial) organisms can be differentially influenced. Probably, when analyzing other compounds in the exudates, such as organic acids, phenolic compounds, other differences between the cultivars can be observed. Further research, under field conditions is necessary to determine the potential to manage the disease by mixing cultivars with different level of resistance.

124

Interaction of banana cultivars with Foc and non-pathogenic F. oxysporum

5.5 Acknowledgements

The authors would like to thank Katia Van Nieuland, Rudy Vergauwen, Tom Struyf, Bruno Jacobson da Silva, Nadia Lemeire and Ilse Delaere for their technical support and Chien-Jui Huang for the inspiring discussions. The study was funded by a PhD scholarship to P.D. from the Fund for Scientific Research-Flanders (FWO).

5.6 Supplementary Material

6 Experiment 1 6 Experiment 2

5 5

4 4

root fresh weight) fresh root 3 3

root fresh weight) fresh root

- -

2 2 log (pg DNA g DNA (pg log

Log (pg DNA g DNA (pg Log 1 1

0 0 Silk Prata Colatina Dwarf Silk Prata Colatina Dwarf Ouro Cavendish Ouro Cavendish

Figure S5.1: The root colonization by pathogenic isolate FocR1 when co-inoculated with three non-pathogenic F. oxysporum isolates (n = 4). Each diamond represents one data point. Diamonds are colored grey when detection was below the detection limit. Silk (AAB), Prata (AAB), Colatina Ouro (AA), Dwarf Cavendish (AAA).

125

Table S5.2: Sugar composition of root exudates in µmol g-1 root of four banana cultivars (Silk (AAB), Prata (AAB), Colatina Ouro (AA) and Dwarf Cavendish (AAA)) when non- inoculated, and one week after inoculation with Foc race 1 (Foc - and Foc +, respectively). The standard deviation in shown between brackets. Different letters show significant differences between cultivar and treatments within one sugar component (Kruskal Wallis, p = 0.05, n = 4).

Treatment total rhamnose arabinose mannose glucose xylose sucrose total root root/shoot weight Silk Foc - 287.87 (44.12) b 76.05 (23.59) bc 55.65 (16.79) b 22.02 (6.66) bc 81.79 (16.50) bc 39.99 (5.79) ab 12.35 (7.47) c 42.30 (9.00) 0.70 (0.16) Foc + 537.86 (136.80) a 137.40 (38.78) a 74.87 (31.32) b 27.68 (5.45) ab 147.76 (24.28) a 72.21 (36.42) a 77.93 (25.66) a 35.65 (10.84) 0.60 (0.08) Prata Foc - 205.18 (79.70) b 29.00 (27.52) c 35.94 (15.72) bc 15.16 (5.18) c 63.91 (18.43) c 28.11 (8.00) bc 33.07 (32.97) bc 81.03 (28.49) 1.03 (0.39) Foc + 278.76 (101.49) b 66.12 (26.61) bc 60.36 (24.69) b 19.36 (8.98) bc 85.41 (27.11) bc 28.08 (8.27) c 19.43 (9.66) c 62.32 (24.21) 1.15 (0.49) Colatina Ouro Foc - 242.69 (63.30) b 48.30 (22.43) bc 46.04 (42.72) bc 14.26 (5.88) c 60.76 (45.10) c 52.59 (19.21) a 20.74 (13.19) c 24.45 (2.89) 0.44 (0.02) Foc + 595.63 (221.85) a 170.57 (102.82) a 139.07 (68.71) a 36.46 (5.86) a 140.70 (67.63) ab 61.44 (29.76) a 47.38 (26.96) ab 22.34 (5.79) 0.41 (0.09) Dwarf Cavendish Foc - 243.55 (90.29) b 85.30 (36.32) b 15.64 (7.84) c 14.58 (8.15) c 80.63 (25.23) c 27.83 (11.27) bc 19.57 (9.01) c 43.34 (5.96) 0.91 (0.15) Foc + 600.25 (128.42) a 133.66 (17.90) a 141.20 (19.46) a 33.75 (7.97) a 183.38 (90.38) a 62.06 (15.87) a 46.20 (8.06) ab 33.45 (7.82) 0.75 (0.08)

126

Interaction of banana cultivars with Foc and non-pathogenic F. oxysporum

30 Foc - a"

25 ab"b" b" * 20

15 a' a ab 10 ab * ab' * 5 ab' a"' b * * b' ab"'b"'b"'

Proportion on total amino acids (%) acids amino total on Proportion 0

30 Foc +

25 a 20

a' 15 ab' ab ab' b b' 10 b a"ab" b"ab"

5 Proportion on total amino acids (%) acids amino total on Proportion

Silk Prata Colatina Ouro Dwarf Cavendish

Figure S5.3: Proportion of different amino acids on the total of amino acids in the root exudates of different banana cultivars, non-inoculated (Foc -), and inoculated with Foc race 1 (Foc +). Different letters show significant differences between the cultivar within the same treatment. When only one value is significantly different from all other values, the different value is indicated with an asterisk (Kruskal Wallis, p = 0.05, n = 4). The error bars show the standard deviation. Silk (AAB), Prata (AAB), Colatina Ouro (AA), Dwarf Cavendish (AAA).

127

Chapter 6

General conclusions and future perspectives

129

Chapter 6

6.1 General conclusions

Farmers’ knowledge and observations are a valuable source of information. This PhD thesis took an interesting farmer’s observation and field situation as starting point to identify potential factors that could contribute to the control of Fusarium wilt on banana. Fusarium wilt is a complicated disease, which requires an integrated approach to be effectively controlled. Its causal agent, F. oxysporum f. sp. cubense is located in soil, where it interacts in a highly complex ecosystem. We tried to increase our understanding on the interaction of Foc and non-pathogenic F. oxysporum with non-susceptible hosts, banana hosts with diverse degrees of resistance and with each other. Below, we resume the main conclusions of this PhD and describe how they can be important for the holistic management of Fusarium wilt on banana.

RQ 1 (Chapter 2): Can differences in Fusarium wilt on the field be attributed to differences in soil disease suppressiveness? If yes, which soil abiotic soil properties, plant community or microbial factors are correlated with Fusarium wilt suppressiveness in soil?

Disease suppressiveness was highest in locations with mixed banana cultivars, less graminoids and higher clay content and pH.

On a farm in Brazil, which was managed as an agroforestry system, banana was affected by Foc race 1. Large variation in disease severity was observed. By greenhouses assays, this variation could be attributed to differences in soil disease suppressiveness. Disease suppression in soil is assumed to be of microbial origin (Alabouvette and Steinberg, 2006). On its turn, the microbial community is the product of interactions with the soil abiotic environments and plant community (Kinkel et al., 2011). To get insight in factors influencing disease suppression, we studied the correlation between disease suppression and the soil abiotic properties, the plant community and the soil microbial community. Three major correlations with soil suppressiveness were found:

130

General conclusion and Future perspectives

- A higher clay content and pH - Lower ground cover by graminoids - A lower density of the susceptible banana cultivar (cv. Maçã) and a high diversity of other banana cultivars.

The increase of soil suppressiveness is a desired trait for farmers. Therefore, the correlating factors described above are potentially useful to manage Fusarium wilt. Soil texture is one of the fixed field characteristics and lies beyond management options. However, it is possible to manage the occurrence of graminoids as weeds or cover crop and the banana cultivars grown. Therefore, based on the two latter correlations, further research questions were formulated (Chapter 4 and 5).

RQ 2 (Chapter 3): How does Foc relate to the local non-pathogenic F. oxysporum population?

Foc isolates with identical sequence type to the local non-pathogenic population were found. Those Foc isolates potentially obtained virulence through horizontal gene transfer.

A collection of pathogenic and non-pathogenic isolates was obtained from soil, graminoids and banana plants of the farm. The pathogenic and co-occurring non-pathogenic population were compared to study if similarities between both could give insight in the origin and nature of the pathogenicity of Foc. Isolates were molecularly characterized and their pathogenicity on banana was evaluated. Two different pathogen types were found: the main group, having an EF-1α/IGS sequence type different from all non-pathogenic isolates. Those Foc isolates probably have been introduced. The second group of Foc isolates shared the same EF- 1α/IGS sequence type as the dominant non-pathogenic population, suggesting their local origin. All pathogens shared the same sequence for SIX1, which was not encountered in non-pathogenic isolates. The same SIX1 sequence has been found previously in Australian Foc (Laurence et al., 2015; Meldrum et al., 2012). Based on the presence of identical SIX1 sequence in all Foc, it can be hypothesized that the Foc isolates with local characteristics acquired pathogenicity through horizontal gene transfer. To our knowledge, this is the first report of potential horizontal gene transfer in Foc.

131

Chapter 6

The report of Foc isolates that potentially arose from horizontal gene transfer is not of pure academic interest. The use of resistant cultivars is often declared to be the sole sustainable solution to manage Foc. Considering the potential of transfer of virulence genes within the FOSC, it is likely that in short time new Foc members with diverse properties can arise. Banana breeding is a slow, challenging process. Therefore, it is important to invest besides resistance breeding in alternative control measures. In addition, the possibility of horizontal gene transfer calls for diagnostic tools based on genes known to be involved in virulence.

RQ 3 (Chapter 4): Do graminoids contribute to pathogen inoculum build-up as symptomless carrier?

Although graminoids are a potential symptomless carrier of Foc, their importance in pathogen inoculum increase depends on the interaction with non-pathogenic F. oxysporum.

F. oxysporum is one of the most common endophytes in plants. In particular, grasslands support a large population of F. oxysporum (Gordon and Martyn, 1997). Therefore, graminoids could be the ideal endophytic carrier of Foc and consequently stimulate the pathogen inoculum density in soil. To explore the potential of graminoids as symptomless carrier of Foc, graminoids were collected on the farm and examined for endophytic Foc isolates. Although from half of the plants F. oxysporum could be isolated, none was pathogenic to banana. Two possible explanations were proposed: (1) Foc has a lower capacity to colonize symptomless hosts compared to non-pathogenic isolates, or, (2) Foc has a lower competitiveness to colonize graminoids when combined with other F. oxysporum isolates. To test these hypotheses, the root and shoot colonization by Foc and three non- pathogenic isolates was analysed for two graminoids: Brachiaria decumbens and Cyperus iria. Foc appeared to have the same capacity to colonize the roots of both graminoids. However, its competitiveness appeared to be lower than the non-pathogens. When co- inoculated, Foc consistently colonized the roots of the graminoids less than the non- pathogenic isolates. The reduced competitiveness of Foc could not be attributed to a reduced growth on abundant carbon sources.

Although our results about the potential field effect of graminoids as symptomless carrier of Foc are not conclusive, they put the concept of symptomless carriers in perspective. Our results clearly demonstrate that the ability of a plant to contain F. oxysporum is not sufficient

132

General conclusion and Future perspectives to be contributing to the increase of pathogen inoculum. Mutual competition between F. oxysporum limited the root colonization by Foc. Despite the ability to contain Foc in their roots, graminoids can rather act as a reservoir of non-pathogenic F. oxysporum with protective abilities. In addition, many other factors of graminoids, such as their contribution to soil organic matter and nutrient competition, need to be taken into consideration to evaluate whether a weed of cover crop will contribute to the disease.

RQ 4 (Chapter 5): Can cultivars with different level of resistance influence disease suppression in soil? Subquestion: How do different cultivars influence the Foc and non-pathogenic F. oxysporum population?

The roots of both the moderately susceptible cultivar and the resistant cultivars showed only limited colonization by Foc and they sustained a restricted Foc population in soil. The non-pathogenic F. oxysporum population in soil was larger after growth of cv. Colatina Ouro compared to the other cultivars.

Members of the F. oxysporum species complex are important players in the suppressiveness of soils. On the one hand, the pathogen density determines the severity of the disease and non-pathogenic F. oxysporum potentially contribute to biocontrol (Alabouvette et al., 2009). Four banana cultivars with different level of resistance to Foc race 1 (Silk, very susceptible, Prata, moderately susceptible, Colatina Ouro and Dwarf Cavendish, both resistant) were studied for their interaction with Foc race 1 and non-pathogenic F. oxysporum. The level of root colonization by Foc race 1 appeared to correspond to the susceptibility of the cultivars. Only in soil planted with the very susceptible cultivar Silk, Foc populations increased. Although cv. Prata showed similar vascular discoloration as cv. Silk, it did not stimulate Foc in soil and Foc colonized its roots only to a limited extent. This suggests that an early defense reaction of cv. Prata, limiting root colonization, can contribute to its partial resistance to Foc race 1. Despite the different root colonization by Foc, non-pathogenic strains colonized the roots of all cultivars equally. Only one cultivar, Colatina Ouro, stimulated non- pathogenic F. oxysporum in soil. This cultivar had the most different amino acid content in its root exudates. Although the root exudates offered no direct explanation for the different soil populations of F. oxysporum, a cultivar-specific shift in sugar and amino acid content was observed after inoculation with Foc. Those shifts might be the result of different defense reactions, but further research is required.

133

Chapter 6

Both the field study as well as the results from chapter 5 support the suggestion that growing a mixture of banana cultivars might contribute to the control of Fusarium wilt. Considering the different impact on the Foc population of cultivars with different resistance level, we assume that growing a mixture will result in a reduced inoculum build-up. The differences in disease observed in the field study however, were not exclusively related to lower inoculum levels. Therefore we investigated if different banana cultivars could also affect other microorganisms, such as non-pathogenic F. oxysporum. The increased non-pathogenic F. oxysporum soil population after growth of cultivar Colatino ouro confirms a broader influence of different cultivars. Also other studies have reported a potential positive effect of cultivar mixtures on the control of Fusarium wilt (Karangwa et al., 2016). However, we acknowledge that cultivar mixtures might be difficult to implement in situations where markets are demanding a sole specific cultivar.

Some suggestions on the nature of resistance in different cultivars have been proposed from the Foc colonization assays and root exudate shifts. Further study of those, especially by comparing cultivars with moderate and full resistance, can contribute to separate the effective from less effective defense reactions. Knowledge of the most effective defense reaction can assist breeders for most effective selection of germplasm.

To conclude this section, the main microbial-microbial-plant interactions studied in this thesis are mapped in Figure 6.1. This figure is based on the general scheme of interplay between the plant and microbial community (Figure 1.5).

134

General conclusion and Future perspectives

Figure 6.1: Overview of the main findings of this PhD. Studied interactions are mapped on the scheme of plant- microbial-microbial interactions. Per chapter a different colour is given. Foc: Fusarium oxysporum f. sp. cubense; Np-Fox: Non-pathogenic F. oxysporum.

6.2 Future perspectives

Most of our general conclusions can be supplemented with several quotes starting with ‘However, further research should be done’. Furthermore, many additional observations have led to new research suggestions. Some future perspectives both on the main and additional observations are discussed below.

 Locations with different level of disease suppressiveness were observed spread over the farm. However, locations with higher level of disease suppression were only observed in the part of the farm managed as agroforestry system (Area A). It could be hypothesized that the enhanced disease suppressiveness is the result of the management under agroforestry. However, as our study was restricted to a single farm, only within farm differences could be compared and no conclusions on the general management could be made. Some evidence is available about damage

135

Chapter 6

reduction of air-borne diseases by agroforestry systems, such as for sigatoka on banana (Norgrove and Hauser, 2013; Schroth, 2000). However, little is known about the effect on soil-borne diseases. For testing the hypothesis that management under agroforestry can induce soil suppressiveness, it is necessary to include more farms managed as agroforestry system and compare them with neighbouring conventional farms. Care should be taken to select farms with the same soil type and same inoculum pressure. Another possibility is to observe several farms that shift from a conventional to an agroforestry system and monitor the impact on disease suppression over time. Finally, the hypothesis could also be tested in a long term field experiment, with block design comparing the effect of the management. Besides the advantage of having more controlled conditions, the required area and labour are a serious drawback.

 As elaborated in the general conclusions, in chapter 3, three Foc isolates that potentially acquired virulence through horizontal gene transfer were described. To confirm the event of horizontal gene transfer, it is necessary to perform a deeper genetic analysis, preferably by full genome sequencing. The threefold comparison of the pathogenic isolates and non-pathogenic isolates with the same EF-1α/IGS, with the other pathogen could be a promising way to gain insight in the pathogenicity related genes in forma specialis cubense.

 All analysed Foc isolates had an identical SIX1 sequence. This sequence was also found previously in Australian Foc isolates (Meldrum et al., 2012; Laurence et al., 2015) and has, to our knowledge, not yet been found in other formae speciales. The exclusive appearance in Foc suggests a potential role of SIX1 in virulence on banana. It might be interesting to study if a Six1 protein can be isolated from the xylem of infected banana plants, like has been done for tomato and F. oxysporum f. sp. lycopersici (Rep et al., 2004).

 The F. oxysporum isolates, which were described in chapter 3, were isolated both from area A and area B of the farm (indicated with PDa and PDb in the name, respectively). It was remarkable that in area A of the farm 36 of 47 non-pathogenic isolates belonged to the dominant group (EF-1α St1-5), while in area B, none of 9 isolates belonged to that group (Table 3.1). The pH was the sole abiotic factor showing significant difference between both areas, nl. pH 5.48 (± 0.62) and 4.52 (± 0.18) in area A and B, respectively. Likewise, Edel et al. (2001) observed, when comparing F. oxysporum populations from different fields, the most different F.

136

General conclusion and Future perspectives

oxysporum population in a field with pH dissimilar to the other examined fields. Besides a different pH, both areas had a different cropping history (Area A: coffee plantation succeeded by agroforestry system with mixture of coffee, banana and trees; Area B: pasture, banana plantation), which could also have a significant influence of the population structure. In area A, banana was grown for a longer period and 92.5% of all non-pathogenic isolates obtained from banana on the field belonged to St1-5. Species specific endophytic selection for this sequence type could be proposed, however, the number of isolations in this study was too low to obtain statistic support. Also no evidence for selective colonization by F. oxysporum could be obtained from the experiment of chapter 5 about the interaction of different banana cultivars with non-pathogenic F. oxysporum. Further research is needed to elucidate factors shaping populations of F. oxysporum. One option to examine the plant selective power on F. oxysporum is to perform more comprehensive sampling from different plant species. To investigate the influence of soil pH on the F. oxysporum community, isolates from soils with different pH can be compared for growth under different pH. Besides a direct influence, it needs to be taken in consideration that the pH will affect other microorganisms which on their turn can affect the F. oxysporum community.

 In chapter 4, we observed that when Foc and non-pathogenic isolates were co- inoculated, Foc was consistently less competitive to colonize the roots of graminoids than the non-pathogenic isolates. These trials were performed with one isolate Foc race 1. However, forma specialis cubense consists of several genetically different isolates (Fourie et al., 2011). Therefore, our results cannot be generalized to all Foc isolates. To test whether the low competitive ability of Foc to colonize roots of non- susceptible hosts is a conservative characteristic of Foc, more Foc isolates should be tested. It can be hypothesized that a long-term close association of Foc and the banana plant could have caused the reduced competitive ability to colonize other plants than banana. When testing more Foc isolates, it would be most interesting to insert Foc isolates that are suspect of having required pathogenicity recently, such as the isolates Ara1, Ara2 and FoxPDa3 together with Foc isolates with long-term pathogenicity. Also the inclusion of other races would be highly interesting. Broader vision on the common characteristics of Foc can be important for the development of consistent effective management.

137

Chapter 6

 In chapter 4, we observed that the two isolates that colonized best the areal parts of C. iria, are less affected in root colonization by competition and showed less dense growth on the different carbon sources. We hypothesized that these isolates have better endophytic properties than the other isolates. It has been investigated whether endophyte populations would be genetically differentiated from soil populations, with variable results (Demers et al., 2015; Edel et al., 1997). From our results it can be suggested that endophytes and saprophytes could rather be differentiated based on phenotypic characteristics, than on genotypic characteristics. We propose that endophytes show better growth on less nutrients and are less influenced by competition for colonization. To test this hypothesis, two different strategies can be followed. Firstly, pot experiments as described in this chapter can be expanded to a larger collection of F. oxysporum isolates. Subsequently, these isolates should be compared for growth characteristics. A second strategy implies the collection of F. oxysporum isolates from areal plant parts and from soil in field and a comparison in growth characteristics between isolates from different origin.

 In chapter 5, four banana cultivars with different levels of resistance to Foc race 1 were evaluated for their influence on the F. oxysporum population. The different impact of the cultivars on the Foc and non-pathogenic population suggests that growing a cultivar mixture might be contributing to the control of Fusarium wilt. However, only separate characteristics were evaluated. Due to time and space constrains no information on the mutual influence was obtained. Ideally, the effect of cultivar mixtures could be tested in participation with farmers willing to plant an area in mixture, next to an unmixed area. Besides Fusarium wilt, many other diseases and pests affect banana. Some studies have reported reduced damage by sigatoka, weevils and nematodes when growing banana in cultivar mixtures (Mulumba et al., 2012; Suarez-Capello and Agama, 2011). Therefore, trials with cultivar mixtures can be additionally evaluated for their impact on other diseases and pest to assess their effectiveness for integrated management. It is important to invest in the development of strategies that focus on the pest complex as a whole, instead of employing individual strategies for control of individual pests or diseases (Staver et al., 2001).

 Four different cultivars were evaluated for their impact on Foc race 1 in chapter 5. In the future, a similar setup could be used to test more cultivars for their contribution to the build-up of Foc race 1 and potentially also other races. In Taiwan, somaclonal cultivars with quantitative resistance to Foc TR4 have been developed. Despite their increased resistance, some scientists are skeptical about their use as they could

138

General conclusion and Future perspectives

contribute to increases in pathogen levels in soil (Dita, 2015). However, today no quantitative data are available. A similar test as in chapter 5 could be used to estimate the impact on TR4 populations of Cavendish cultivars, their somaclonal variants and other hybrids developed for their resistance to Foc TR4.

 After inoculation with Foc race 1, a different shift in the sugar and amino acid content of the root exudates of the different cultivars was observed. Since defense reactions have a big impact on the primary metabolism of a plant, these shifts might be the result of different defense mechanisms. However, the collection and analysis of the root exudates has been performed only once and should be repeated to confirm the results. Analysis of enzymatic activity, microscopical imaging or transcriptome analysis can be used to link defense mechanisms to the shift in metabolites in the root exudates after inoculation. For instance, the decrease of phenylalanine in the root exudate of cv. Prata could be explained by increase in PAL activity. Most remarkable was the increase of arabinose in the root exudate of both resistant cultivars, which could be explained by a change in the pectins. Increased understanding of the effectiveness of different defense reactions in banana obtained by comparing banana plants with different level of resistance can help breeders to find the most effective QRL. After confirmation that defense reactions are easily visualized in root exudates, the analysis of root exudates might offer a relatively fast and easy alternative to evaluate germplasm.

 While testing non-pathogenic F. oxysporum for their protective potential, typical vascular discoloration was observed in several plants which were not inoculated with the pathogen. Inoculation protocols were followed with care and transmission via water, soil, tools or infected suckers could be excluded. During the experiment, there was a serious outbreak of fungus gnats (Sciaridae) in the growth room. Those insects have a larval stage which feeds on plant roots and fungi and a flying adult stage. We hypothesized that the Sciarids acted as vector of Foc. Foc race 1 could be isolated from adult fungus gnats (Heyman, 2015). Despite the highly artificial situation of this observation, in a growth room with high levels of Sciarids, transmission by fungus gnats can be a way of dispersal of Foc in nature. Fungus gnats are one of the most abundant insects in tropical soils (Nielsen and Nielsen, 2004) and flying adults can, with the support of wind, travel several hundred meters. This could explain the spread of TR4 beyond neigbouring plants in banana plantations in Australia which are under strict quarantine measures (Meldrum et al., 2013). The importance of Sciarids in the spread of F. oxysporum has already been assessed in the greenhouse, where

139

Chapter 6

Sciarids can occur as pest (Gillespie and Menzies, 1993; Scarlett et al., 2014). However, further field research is necessary to estimate the potential and importance of Sciarids as Foc vector.

I hope that this work might be an inspiring source for others. You’re kindly invited to dig further in the world of F. oxysporum and banana and allowed to cross contaminate your work with these suggestions.

140

References

Abadie C, Edel V, Alabouvette C, 1997. Soil suppressiveness to Fusarium wilt: influence of a cover-plant on density and diversity of Fusarium populations. Soil Biology and Biochemistry 30, 643-649.

Aguilar EA, Turner DW, Sivasithamparam K, 2000. Fusarium oxysporum f. sp. cubense inoculation and hypoxia alter peroxidase and phenylalanine ammonia lyase activities in nodal roots of banana cultivars (Musa sp.) differing in their susceptibility to Fusarium wilt. Australian Journal of Botany 48, 589–596.

Alabouvette C, 1986. Fusarium-wilt suppressive soils from the Chateaurenard region: review of a 140-year study. Agronomie 6, 273-284.

Alabouvette, C 1999. Fusarium wilt suppressive soils: an example of disease-suppressive soils. Australasian Plant Pathology 28, 57-64.

Alabouvette C, Olivain C, Migheli Q, Steinberg C, 2009. Microbiological control of soil-borne phytopathogenic fungi with special emphasis on wilt-inducing Fusarium oxysporum. New Phytologist 184, 529–544.

Alabouvette C, Steinberg C, 2006. The soil as a reservoir for antagonists to plant diseases. In: Eilenberg J, Hokkanen HMT, eds. An ecological and societal approach to biological control. Dordrecht, the Netherlands: Springer, 123–144.

Alves ÉJ, 2009. A cultura da banana: aspectos técnicos, socioeconômicos e agroindustriais. Brasília: Embrapa-SPI/Cruz das Almas: Embrapa-CNPME, 585pp.

Altinok HH, 2013. Fusarium species isolated from common weeds in eggplant fields and symptomless hosts of Fusarium oxysporum f. sp. melongenae in Turkey. Journal of Phytopathology 161, 335-340.

Altmann F, 1992. Determination of amino sugars and amino acids in glycoconjugates using precolumn derivatization with o-phthalaldehyde. Analytical Biochemistry 204, 215- 219.

Amir H, Alabouvette C, 1993. Involvement of soil abiotic factors in the mechanisms of soil suppressiveness to fusarium wilts. Soil Biology and Biochemistry 25, 157-164.

Aoki T, O’Donnell K, Geiser DM, 2014. Systematics of key phytopathogenic Fusarium species: current status and future challenges. Journal of General Plant Pathology 80, 189–201.

Bais HP, Weir TL, Perry LG, Gilroy S, Vivanco JM, 2006. The role of root exudates in rhizosphere interactions with plants and other organisms. Annual Review of Plant Biology 57, 233–266.

141

Baker KF, Cook RJ, 1974. Biological control of plant pathogens. W. H. Freeman, San Fransico, 433 pp.

Balmas V, Migheli Q, Scherm B, Garau P, O’Donnell K, Ceccherelli G, Kang S, Geiser D M, 2010. Multilocus phylogenetics show high levels of endemic Fusaria inhabiting Sardinian soils (Tyrrhenian Islands). Mycologia 102, 803-812.

Belgrove A, Steinberg C, Viljoen A, 2011. Evaluation of nonpathogenic Fusarium oxysporum and Pseudomonas fluorescens for Panama disease control. Plant Disease 95, 951- 959.

Benhamou N, Grenier J, Chrispeels M J, 1991. Accumulation of fl-Fructosidase in the cell walls of tomato roots following infection by a fungal wilt pathogen. Plant Physiology 97, 739-750.

Benhamou N, Garand C, Goulet A, 2002. Ability of non-pathogenic Fusarium oxysporum Fo47 to induce resistance against Pythium ultimum infection in cucumber. Applied and Environmental Microbiology 68, 4044–4060.

Ben-Yephet Y, Reuveni M, Genizi A, 1994. Effects of inoculum depth and density on Fusarium wilt in carnations. Phytopathology 84, 1393–1398.

Berger S, Sinha AK, Roitsch T, 2007. Plant physiology meets phytopathology: plant primary metabolism and plant–pathogen interactions. Journal of Experimental Botany 58, 4019–4026.

Bertin C, Yang X, Weston LA, 2003. The role of root exudates and allelochemicals in the rhizosphere. Plant and Soil 256, 67–83.

Blouin M, Hodson ME, Delgado EA, Baker G, Brussaard L, Butt KR, Dai J, Dendooven L, Peres G, Tondoh JE, Cluzeau D, Brun J-J, 2013. A review of earthworm impact on soil function and ecosystem services. European Journal of Soil Science 64, 161-182.

Bolton MD, 2009. Primary metabolism and plant defense—fuel for the fire. Molecular Plant- Microbe Interactions 22, 487–497.

Borges AL, da Silva Souza L, 2004. O cultiva da bananeira. Embrapa Mandioca e Fruticultura, Cruz das Almas, 182 pp.

Borges AL, da Silva AL, Batista D da C, Moreira, FRB, Flori, JE, Oliveira JE de M, Araujo, JLP, Pinto, JM, Castro, JM da C, Moura MSB de, Azoubel PM, Cunha TJF, Silva, S de O, Cordeiro, ZJM 2009. Sistema de Produção da Bananeira Irrigada. Embrapa, Brazil, 115pp.

Broeckling CD, Broz AK, Bergelson J, Manter DK, Vivanco JM, 2008. Root exudates regulate soil fungal community composition and diversity. Applied and Environmental Microbiology 74, 738-744.

Butler D, 2013. Fungus threatens top banana. Nature 504, 195-196.

Buxton EW, 1962. Root exudates from banana and their relationship to strains of the Fusarium causing Panama wilt. Annals of Applied Biology 50, 269 - 282.

Canero DC, Roncero MIG, 2008. Functional analyses of laccase genes from Fusarium oxysporum. Phytopathology 98, 509–518.

142

Caraguel CGB, Stryhn H, Gagne N, Dohoo IR, Hammell KL, 2011. Selection of a cutoff value for real-time polymerase chain reaction results to fit a diagnostic purpose: analytical and epidemiologic approaches. Journal of Veterinary Diagnostic Investigation 23, 2– 15.

Cardoso IM, Guijt I, Franco FS, Carvalho AF, Ferreira Neto PS, 2001. Continual learning for agroforestry system design: university, NGO and farmer partnership in Minas Gerais, Brazil. Agroforestry Systems 69, 235–257.

Carlier J, De Waele D, Escalant JV, 2003. Global evaluation of Musa germplasm for resistance to Fusarium wilt, Mycosphaerella leaf spot diseases and nematodes. Performance evaluation (A. Vézina and C. Picq, eds). INIBAP Technical Guidelines 7. The International Network for the Improvement of Banana and Plantain, Montpellier, France.

Chakrabarti A, Rep M, Wang B, Ashton A, Dodds P, Ellis J, 2011. Variation in potential effector genes distinguishing Australian and non-Australian isolates of the cotton wilt pathogen Fusarium oxysporum f.sp. vasinfectum. Plant Pathology 60, 232–243.

Chaparro JM, Sheflin AM, Manter DK, Vivanco JM, 2012. Manipulating the soil microbiome to increase soil health and plant fertility. Biology and Fertility of Soils 48, 489–499.

Cook R, Baker KF, 1983. The nature and practice of biological control of plant pathogens. American Phytopathological Society, St Paul, Minnesota, 539 pp.

Costa SN, Bragan CAD, Ribeiro LR, Amorim EP, Oliveira SAS, Dita MA, Laranjeira FF, Haddad F, 2015. Genetic structure of Fusarium oxysporum f. sp. cubense in different regions from Brazil. Plant Pathology 64, 137–146.

Cunha CMS, Hinz RH, Pereira A, Tcacenco FA, Stadnik MJ, 2015. Aggressiveness and genetic diversity of Fusarium oxysporum f. sp. cubense from Santa Catarina, southern Brazil. Tropical Plant Pathology 5, 326 – 334.

Daniells J, Lindsay S, 2016. TR4 as a driver of agroecological approaches in banana production. http://www.promusa.org/blogpost468-TR4-as-a-driver-of-agroecological- approaches-in-banana-production, accessed on dec 20th 2016.

De Ascensao ARFDC, Dubery IA, 2003. Soluble and wall-bound phenolics and phenolic polymers in Musa acuminata roots exposed to elicitors from Fusarium oxysporum f.sp. cubense. Phytochemistry 63, 679–686.

Debode J ,Van Hemelrijck W, Creemers P, Maes M, 2013. Effect of fungicides on epiphytic yeasts associated with strawberry. MicrobiologyOpen 2, 482–491.

Défago G, Kern H, Sedlar L, 1983. Genetic analysis of tomatine insensitivity, sterol content and pathogenicity for green tomato fruits in mutants of Fusarium solani. Physiological and Molecular Plant Pathology 22, 39–43.

De Langhe E, Vrydaghs L, de Maret P, Perrier X, Denham T, 2009. Why bananas matter: An introduction to the history of banana domestication. Ethnobotany Research and Applications 7, 165-177.

143

Demers JE, Gugino BK, Jiménez-Gasco MdM, 2015. Highly diverse endophytic and soil Fusarium oxysporum populations associated with field-grown tomato plants. Applied Environmental Microbiology 81, 81–90. de Oliveira e Silva S, Teixeira Souza Junior M, Alves ÉJ, Silva Silveira JR, Bezerra Lima M, 2001. Banana breeding program at Embrapa. Crop Breeding and Applied Biotechnology 1, 399-436. de Sain M, Rep M, 2015. The role of pathogen-secreted proteins in fungal vascular wilt diseases. International Journal of Molecular Sciences 16, 23970-23993.

Ding Z, Li M, Sun F, Xi P, Sun L, Zhang L, Jiang Z, 2015. Mitogen-activated protein kinases are associated with the regulation of physiological traits and virulence in Fusarium oxysporum f. sp. cubense. PLoS One 10 (4). di Pietro A, Madrid MP, Caracuel Z, Delgado-Jarana J, Roncero MIG, 2003. Fusarium oxysporum: exploring the molecular arsenal of a vascular wilt fungus. Molecular Plant Pathology 4, 315 – 325.

Dita MA, Waalwijk C, Buddenhagen IW, Souza JrMT, Kema GHJ, 2010. A molecular diagnostic for tropical race 4 of the banana fusarium wilt pathogen. Plant Pathology 59, 348-357.

Dita M, 2015. Why screening protocols matter. ProMusa newsletter. http://www.promusa.org/blogpost406-Why-screening-protocols-matter. Accessed on Feb 12, 2017.

Domínguez J, Negrín MA, Rodríguez CM, 2001. Aggregate water-stability, particle-size and soil solution properties in conducive and suppressive soils to Fusarium wilt of banana from Canary Islands (Spain). Soil Biology and Biochemistry 33, 449-455.

Dong X, Ling N, Wang M, Shen Q, Guo S, 2012. Fusaric acid is a crucial factor in the disturbance of leaf water imbalance in Fusarium-infected banana plants. Plant Physiology and Biochemistry 60, 171-179.

Doornbos R, Loon L, Bakker P, 2012. Impact of root exudates and plant defense signaling on bacterial communities in the rhizosphere. A review. Agronomy for Sustainable Development 32, 227–243.

Dorel M, Damour G, Salmon F, Ricci S, de Lapeyre L, 2016. What variety for sustainable banana cropping systems: Cavendish cultivars or Black Sigatoka resistant hybrids? Management of banana functional traits. Oral presentation at the Promusa conference, Montpellier, France. http://www.promusa.org/article142-2016-France- symposium, accessed on Feb 9, 2017.

Edel V, Steinberg C, Gautheron N, Alabouvette C, 1997. Populations of nonpathogenic Fusarium oxysporum associated with roots of four plant species compared to soilborne populations. Phytopathology 87, 693-697.

Edel V, Steinberg C, Gautheron N, Recorbet G, Alabouvette C, 2001. Genetic diversity of Fusarium oxysporum populations isolated from different soils in France. FEMS Microbiology Ecology 36, 61-71.

144

Elmer WH, Lacy WL, 1987. Effects of crop residues and colonization of plant tissues on propagule survival and soil populations of Fusarium oxysporum f. sp. apii Race 2. Phytopathology 77, 381-387.

Endah R, Beyene G, Kiggundu A, Van den Berg N, Schluter U, Kunert K, Chikwamba R, 2008. Elicitor and Fusarium-induced expression of NPR1-like genes in banana. Plant Physiology and Biochemistry 46, 1007–1014.

FAO data base: http://www.fao.org/economic/worldbananaforum/statistics/en/, accessed on Dec 16th 2016.

FAO, 2014. Banana Market Review and Banana Statistics 2012-2013. Market and Policy Analyses of Raw Materials, Horticulture and Tropical (RAMHOT) Products Team. http://www.fao.org/docrep/019/i3627e/i3627e.pdf, accessed on Feb 28th, 2017.

Fernandes JS, Angelo PCS, Cruz JC, Santos JMM, Sousa NR, Silva GF, 2016. Post- transcriptional silencing of the SGE1 gene induced by a dsRNA hairpin in Fusarium oxysporum f. sp cubense, the causal agent of Panama disease. Genetics and Molecular Research 15 (2).

Fitzpatrick, DA, 2011. Horizontal gene transfer in fungi. FEMS Microbiology letters 329, 1-8.

Forsyth M, Smith LJ, Aitken ABE, 2006. Identification and characterization of non-pathogenic Fusarium oxysporum capable of increasing and decreasing Fusarium wilt severity. Mycological Research 110, 929 – 935.

Fourie G, Steenkamp ET, Gordon TR, Viljoen A, 2009. Evolutionary relationships among the Fusarium oxysporum f. sp. cubense vegetative compatibility groups. Applied and Environmental microbiology 75, 4770–4781.

Fourie G, Steenkamp ET, Ploetz RC, Gordon TR, Viljoen A, 2011. Current status of the taxonomic position of Fusarium oxysporum formae specialis cubense within the Fusarium oxysporum complex. Infection, Genetics and Evolution 11, 533–542.

Fraser-Smith S, Czislowskia E, Meldrum RA, Zandera M, O’Neill W, Balalie GR, Aitken, EAB, 2014. Sequence variation in the putative effector gene SIX8 facilitates molecular differentiation of Fusarium oxysporum f. sp. cubense. Plant Pathology 63, 1044 – 1052.

Furtado EL, Bueno C J, de Oliveira ALM, Menten JO, Malavolta E, 2009. Relações entre ocorrência do Mal-de-Panama em bananeira da cv. Nanicão e nutrientes no solo e nas folhas. Tropical Plant Pathology 34, 211-215.

Garbeva P, van Veen JA, van Elsas JD, 2004. Microbial diversity in soil: selection of microbial populations by plant and soil type and implications for disease suppressiveness. Annual Review of Phytopathology 42, 243-270.

Garcia FA, Ordonez N, Konkol J, Al Qasem M, Naser Z, Abdel wali M, Salem NM, Waalwijk C, Ploetz RC, Kema G, 2014. First report of Fusarium oxysporum f. sp. cubense tropical race 4 associated with Panama disease of banana outside Southeast Asia. Plant Disease 98, 694.

Garrett, 1970. Pathogenic root-infecting fungi. Cambridge university press, London, United Kindom, 294 pp.

145

Gawehns F, Ma L, Bruning O, Houterman PM, Boeren S, Cornelissen BJC, Rep M, Takken, FLW, 2015. The effector repertoire of Fusarium oxysporum determines the tomato xylem proteome composition following infection. Frontier in Plant Science 6, 967.

Geiser DM, Jiménez-Gasco, MdM, Kang S, Makalowska I, Veeraraghavan N, Ward TJ, Zhang N, Kuldau GA, O’Donnell K, 2004. FUSARIUM-ID v. 1.0: A DNA sequence database for identifying Fusarium. European Journal of Plant Pathology 110, 473– 479.

Gillespie DR, Menzies JG, 1993. Fungus gnats vector Fusarium oxysporum f. sp. radices- lycopersici. Annals of Applied Biology 123, 539-544.

Gordon TR, Okamoto D, Jacobson DJ, 1989. Colonization of muskmelon and nonsusceptible crops by Fusarium oxysporum f. sp. melonis and other species of Fusarium. Phytopathology 79, 1095-1100.

Gordon TR, Martyn RD, 1997. The evolutionary biology of Fusarium oxysporum. Annual Review of Phytopathology 35, 111–128.

Gómez-Doñate M, Casanovas-Massana A, Muniesa M, Blanch AR, 2016. Development of new host-specific Bacteroides qPCRs for the identification of fecal contamination sources in water. MicrobiologyOpen 5, 83–94.

Grayston SJ, Wang S, Campbell CD, Edwards AC, 1998. Selective influence of plant species on microbial diversity in the rhizosphere. Soil Biology and Biochemistry 30, 369-78.

Guo L, Han L, Yang L, Zeng H, Fan D, Zhu Y, Feng Y, Wang G, Peng C, Jiang X, Zhou D, Ni P, Liang C, Liu L, Wang J, Mao C, Fang X, Peng M, Huang J, 2014. Genome and transcriptome analysis of the fungal pathogen Fusarium oxysporum f. sp. cubense causing banana vascular wilt disease. PlosOne 9 (4).

Guo L, Yang L, Liang C, Wang G, Dai Q, Huang J, 2015. Differential colonization patterns of bananas (Musa spp.) by physiological race 1 and race 4 Isolates of Fusarium oxysporum f.sp. cubense. Journal of Phytopathology 163, 807–817.

Hage-Ahmed K, Moyses A, Voglgruber A, Hadacek F, Steinkellner S, 2013. Alterations in root exudation of intercropped tomato mediated by the arbuscular mycorrhizal fungus Glomus mosseae and the soilborne pathogen Fusarium oxysporum f.sp. lycopersici. Journal of Phytopathology 161, 763–773.

Hall V, O’Neill GL, Magee JT, Duerden BI, 1999. Development of amplified 16S ribosomal DNA restriction analysis for identification of Actinomyces species and comparison with pyrolysis-mass spectrometry and conventional biochemical tests. Journal of Clinical Microbiology 37, 2255-2261.

Hallmann J, Quadt-Hallmann A, Mahaffee WF, Kloepper JW, 1997. Bacterial endophytes in agricultural crops. Canadian Journal of Microbiology 43, 895–914.

Happi EmagaT, Robert C, Ronkart SN, Wathelet B, Paquot M, 2008. Dietary fibre components and pectin chemical features of peels during ripening in banana and plantain varieties. Bioresource Technology 99, 4346-4354.

Hao WY , Ren LX, Ran W, Shen QR, 2010. Allelopathic effects of root exudates from watermelon and rice plants on Fusarium oxysporum f.sp. niveum. Plant and Soil 336, 485–497.

146

Hardoim PR, van Overbeek LS, van Elsas JD, 2008. Properties of bacterial endophytes and their proposed role in plant growth. Trends in Microbiology 16, 463-471.

Helbig JB, Carroll RB, 1984. Dicotyledonous weeds as a source of Fusarium oxysporum pathogenic on soybean. Plant Disease 68, 694-696.

Hennessy C, Walduck G, Daly A, Padovan A, 2005. Weed hosts of Fusarium oxysporum f. sp. cubense tropical race 4 in northern Australia. Australasian Plant Pathology 34, 115-117.

Hermanto C, Djatnika, EI, Emilda, Mujiman D, Subhana, 2012. Pre-planting treatments for management of banana fusarium wilt. ARPN Journal of Agricultural and Biological Science 7, 260-265.

Hewett P, Ganser GH, 2007. A comparison of several methods for analyzing censored data. Annals of Occupational Hygiene 51, 611–632.

Heyman L, 2015. Non-pathogenic Fusarium oxysporum populations as drivers of soil suppressiveness to Fusarium wilt in banana. Master thesis. Ghent University, Ghent Belgium.

Höper H, Steinberg C, Alabouvette C, 1995. Involvement of clay type and pH in mechanisms of soil suppressiveness to Fusarium wilt of flax. Soil Biology and Biochemistry 27, 955-967.

Hopkins DL, Larkin RP, Elmstrom GW, 1987. Cultivar-specific induction of soil suppressiveness to Fusarium wilt of watermelon. Phytopathology 77, 607-611.

Houterman PM, Cornelissen B, Rep M, 2008. Suppression of plant resistance gene-based immunity by a fungal effector. PLOS Pathogens 4, e1000061.

Houterman PM, Ma L, van Ooijen G, de Vroomen MJ, Cornelissen BJ, Takken FL, Rep M, 2009. The effector protein Avr2 of the xylem-colonizing fungus Fusarium oxysporum activates the tomato resistance protein I-2 intracellularly. The Plant Journal 58, 970– 978.

Huang BZ, Xu LB, Molina AB, 2005. Preliminary evaluation of IMTP-III varieties and local cultivars against Fusarium wilt disease in southern China. Proceedings of the 3rd BAPNET Steering Committee meeting held in Guangzhou 23-26 November 2004. A.B. Molina, V.N. Roa, I. Van den Bergh and K.H. Borromeo (eds). Advancing Banana and Plantain R&D in Asia and the Pacific, Vol. 13:187-191. INIBAP, Los Baños, Laguna, Philippines.

Huang B, Xu Y, Wu Y, Zhang S, Chen X, 2008. Effect of root exudates of different resistant varieties of cucumber on fusarium wilt and preliminary studies on their resistance mechanism. Frontiers of Agriculture in China 2, 61-65.

Huang YH, Wang RC, Li CH, Zuo CW, Wei YR, Zhang L, Yi GJ, 2012. Control of Fusarium wilt in banana with Chinese leek. European Journal of Plant Pathology 134, 87–95.

Huang L-F, Song L-X, Xia X-J, Mao W-H, Shi K, Zhou Y-H, Yu J-Q, 2013. Plant-soil feedbacks and soil sickness: from mechanisms to application in agriculture. Journal of Chemical Ecology 39, 232–242.

147

Huang X-F, Chaparro JM, Reardon KF, Zhang R, Shen Q, Vivanco JM, 2014. Rhizosphere interactions: root exudates, microbes, and microbial communities. Botany 92, 267 – 275.

Imazaki I, Kurahashi M, Iida Y, Tsuge T, 2007. Fow2, a Zn(II)2Cys6-type transcription regulator, controls plant infection of the vascular wilt fungus Fusarium oxysporum. Molecular Microbiology 63, 737– 753.

Inderjit, Weiner J, 2001. Plant allelochemical interference or soil chemical ecology? Perspectives in Plant Ecology, Evolution and Systematics 4: 3-12.

Janvier C, Villeneuve F, Alabouvette C, Edel-Hermann V, Mateille T, Steinberg C, 2007. Soil health through soil disease suppression: Which strategy from descriptors to indicators? Soil Biology and Biochemistry 39, 1-23.

Jin X, Feng D, Wang H, Wang J, 2007. A novel tissue-specific plantain beta-1,3-glucanase gene that is regulated in response to infection by Fusarium oxysporum f.sp. cubense. Biotechnology Letters 29, 1431–1437.

Jorgensen SE, Nielsen SN, 1996. Application of ecological engineering principles in agriculture. Ecological engineering 7, 373-381.

Karangwa P, Blomme G, Beed F, Niyongere C, Viljoen A, 2016. The distribution and incidence of banana Fusarium wilt in subsistence farming systems in east and central Africa. Crop Protection 84, 132-140.

Kashiwa T, Inami K, Fujinaga M, Ogiso H, Yoshida T, Teraoka T, Arie T, 2013. An avirulence gene homologue in the tomato wilt fungus Fusarium oxysporum f. sp. lycopersici race 1 functions as a virulence gene in the cabbage yellows fungus F. oxysporum f. sp. conglutinans. Journal of General Plant Pathology 79, 412–421.

Katan J, 1971. Symptomless carriers of the tomato Fusarium wilt pathogen. Phytopathology 61, 1213-1217.

Kavino M, Kumar N, Damodaran T, Harish S, Saravanakumar D, 2007. Biochemical markers as a useful tool for the early identification of Fusarium oxysporum f.sp. cubense, race1 resistance banana clones. Archives of Phytopathology and Plant Protection 42, 1069–1078.

Kennedy PG, Higgins LM, Rogers RH, Weber MG, 2011. Colonization-competition tradeoffs as a mechanism driving successional dynamics in ectomycorrhizal fungal communities. PLoS ONE 6 (9).

Kinkel LL, Bakker MG, Schlatter DC, 2011. A coevolutionary framework for managing disease-suppressive soils. Annual Review of Phytopathology 49, 47-67.

Koenig RL, Ploetz RC, Kistler HC, 1997. Fusarium oxysporum f. sp. cubense consists of a small number of divergent and globally distributed clonal lineages. Phytopathology 87, 915-923.

Komada H, 1975. Development of a selective medium for quantitative isolation of Fusarium oxysporum from natural soil. Review of Plant Protection Research 8, 114-125.

Kurtz A, 2010. Endophytic Fusarium oxysporum: Phylogeny and induced defense responses in banana plants against Radopholus similis. PhD thesis. Institut für

148

Nutzpflanzenwissenschaften und Ressourcenschutz der Rheinischen Friedrich- Wilhelms-Universität Bonn.

Larkin RP, Fravel DR, 1999. Mechanims of action and dose-response relationships governing biological control of Fusarium wilt of tomato by non-pathogenic Fusarium spp. Phytopathology 89, 1152–1161.

Larkin RP, Hopkins DL, Martin FN, 1993. Ecology of Fusarium oxysporum f. sp. niveum in soils suppressive and conducive to fusarium wilt of watermelon. Phytopathology 83, 1105-1116.

Laurence MH, Summerell BA, Burgess LW, Liew ECY, 2014. Genealogical concordance phylogenetic species recognition in the Fusarium oxysporum species complex. Fungal Biology 118, 374 – 384.

Laurence MH, Summerell BA, Liew ECY, 2015. Fusarium oxysporum f. sp. canariensis: evidence for horizontal gene transfer of putative pathogenicity genes. Plant Pathology 64, 1068-1075.

Lavelle P, Spain A, Blouin M, Brown G, Decaens T, Grimaldi M, Jimenez JJ, McKey D, Mathieu J, Velasquez E, Zangerle A, 2016. Ecosystem engineers in a self-organized soil: A review of concepts and future research questions. Soil Science 181, 91-109.

Leoni C, de Vries M, ter Braak CJF, van Bruggen AHC, Rossing WAH, 2013. Fusarium oxysporum f.sp. cepae dynamics: in-plant multiplication and crop sequence simulations. European Journal of Plant Pathology 137, 545–561.

Lescot T, 2015. Genetic diversity of the banana. Fruitrop 231:98-102.

Levy-Booth DJ, Campbell RG, Gulden RH, Hart MM, Powell JR, Klironomos JN, Pauls KP, Swanton CJ, Trevors JT, Dunfield KE, 2007. Cycling of extracellular DNA in the soil environment. Soil Biology and Biochemistry 39, 2977–2991.

Li XG, Liu B, Heia S, Liu DD, Han ZM, Zhou KX, Cui JJ, Luo JY, Zheng YP, 2009. The effect of root exudates from two transgenic insect-resistant cotton lines on the growth of Fusarium oxysporum. Transgenic Research 18, 757–767.

Li C, Chen S, Zuo C, Sun Q, Ye Q, Yi G, Huang B, 2011a. The use of GFP-transformed isolates to study infection of banana with Fusarium oxysporum f. sp. cubense race 4. European Journal of Plant Pathology 131, 327–340.

Li WM, Qian CM, Mo YW, Hu YL, Xie JH, 2011b. Tolerance of banana for fusarium wilt is associated with early H2O2 accumulation in the roots. African Journal of Biotechnology 10, 11378–11387.

Li C, Deng GM, Yang J, Viljoen A, Jin Y, Kuang RB, Zuo CW, Lv ZC, Yang QS, Sheng O, Wei YR, Hu CH, Dong T, Yi GJ, 2012. Transcriptome profiling of resistant and susceptible Cavendish banana roots following inoculation with Fusarium oxysporum f. sp. cubense tropical race 4. BMC Genomics 13, 374.

Li C, Shao J, Wang Y, Li W, Guo D, Yan B, Xia Y, Peng M, 2013. Analysis of banana transcriptome and global gene expression profiles in banana roots in response to infection by race 1 and tropical race 4 of Fusarium oxysporum f. sp. cubense. BMC Genomics 14, 851.

149

Li W, Ge X, Wu W, Wang W, Hu Y, Mo Y, Sun D, Shi S, Xie J, 2015. Identification of defense-related genes in banana roots infected by Fusarium oxysporum f. sp. cubense tropical race 4. Euphytica 205, 837–849.

Lievens B, Houterman PM, Rep M, 2009. Effector gene screening allows unambiguous identification of Fusarium oxysporum f. sp. lycopersici races and discrimination from other formae speciales. FEMS Microbiology Letters 300, 201-215.

Lori G, Edel-Hermann V, Gautheron N, Alabouvette C, 2004. Genetic diversity of pathogenic and nonpathogenic populations of Fusarium oxysporum isolated from carnation fields in Argentina. Phytopathology 94, 661 – 668.

Louvet J, Rouxel F, Alabouvette C, 1976. Recherches sur la résistance des sols aux maladies. I - Mise en évidence de la nature microbiologique de la résistance d’un sol au développement de la fusariose vasculaire du melon. Annales de Phytopathologie 8, 425–436.

Lugtenberg B, Kamilova F, 2009. Plant-growth-promoting rhizobacteria. Annual Review of Microbiology 63, 541–556.

Ma LJ, van der Does HC, Borkovich KA, Coleman JJ, Daboussi MJ, di Pietro A, Dufresne M, Freitag M, Grabherr M, Henrissat B, Houterman PM, Kang S, et al. 2010. Comparative genomics reveals mobile pathogenicity chromosomes in Fusarium. Nature 464, 367 – 373.

Ma LJ, 2014. Horizontal chromosome transfer and rational strategies to manage Fusarium vascular wilt diseases. Molecular Plant Pathology 15, 763-766.

Madrid MP, di Pietro, A, Roncero MIG, 2003. Class V chitin synthase determines pathogenesis in the vascular wilt fungus Fusarium oxysporum and mediates resistance to plant defense compounds. Molecular Microbiology 47, 257–266.

Mak C, Mohamed AA. Liew KW, Ho YW, 2004. Early screening technique for Fusarium wilt resistance in banana micropropagated plants, in: Jain, S. M., Swennen, R. (Eds.), Banana Improvement: cellular, molecular biology and induced mutations. Science Publishers, Inc. Enfield, USA, pp. 219–227.

Malcolm GM, Kuldau GA, Gugino BK, Jiménez-Gasco MdM, 2013. Hidden host plant associations of soilborne fungal pathogens: An ecological perspective. Phytopathology 103, 538-544.

Mandeel Q, Baker R, 1991. Mechanisms involved in biological control of Fusarium wilt of cucumber with strains of nonpathogenic Fusarium oxysporum. Phytopathology 81, 462-469.

Mangla S, Callaway RM, 2008. Exotic invasive plant accumulates native soil pathogens which inhibit native plants. Journal of Ecology 96, 58 - 67.

Mano H, Tanaka F, Nakamura C, Kaga H, Morisaki H, 2007. Culturable endophytic bacterial flora of the maturing leaves and roots of rice plants (Oryza sativa) cultivated in a paddy field. Microbes and environment 22, 175-185.

Meldrum RA, Fraser-Smith S, Tran-Nguyen LTT, Daly AM, Aitken EAB, 2012. Presence of putative pathogenicity genes in isolates of Fusarium oxysporum f. sp. cubense from Australia. Australasian Plant Pathology 41, 551-557.

150

Meldrum RA, Daly AM, Tran-Nguyen LTT, Aitken EAB, 2013. Are banana weevil borers a vector in spreading Fusarium oxysporum f. sp. cubense tropical race 4 in banana plantations? Australasian Plant Pathology 42, 543-549.

Michielse CB, Rep M, 2009. Pathogen profile update: Fusarium oxysporum. Molecular Plant Pathology 10, 311-324.

Michielse CB, van Wijk R, Reijnen L, Cornelissen BJ, Rep M, 2009a. Insight into the molecular requirements for pathogenicity of Fusarium oxysporum f. sp. lycopersici through large-scale insertional mutagenesis. Genome Biology 10, R4.

Michielse CB, van Wijk R, Reijnen L, Manders EMM, Boas S, Olivain C, Alabouvette C, Rep M, 2009b. The nuclear protein Sge1 of Fusarium oxysporum is required for parasitic growth. PLoS Pathogens 5 (10).

Miller RNG, Bertioli DJ, Baurens FC, Santos CMR, Alves PC, Martins NF, Togawa RC, Souza MT, 2008. Analysis of non-TIR NBS-LRR resistance gene analogs in Musa acuminata Colla: isolation, RFLP marker development, and physical mapping. BMC Plant Biology 8, 15.

Minerdi D, Moretti M, Gilardi G, Barberio C, Gullino ML, Garibaldi A, 2008. Bacterial ectosymbionts and virulence silencing in a Fusarium oxysporum strain. Environmental Microbiology 10, 1725–1741.

MGIS, 2017. https://www.crop-diversity.org/mgis/, accessed on May 16, 2017.

Moberg M, Striffler S, 2003. Banana wars : power, production, and history in the Americas. Durham (N.C.) : Duke university press, 364 pp.

Moe LA, 2013. Amino acids in the rhizosphere: From plants to microbes. American Journal of Botany 100, 1692–1705.

Moore NY, Bentley S, Pegg KG, Jones DR, 1995. Fusarium wilt of Banana. Musa Disease Fact Sheet N° 5. INIBAP, Montpellier, France.

Moreira RS, 1999. Banana – Teoria e prática de cultivo, 2.ed.: Fundação Cargill, São Paulo, 657 pp.

Mulumba JW, Nankya R, Adokorach J, Kiwuka C, Fadda C, De Santis P, Jarvis DI, 2012. A risk-minimizing argument for traditional crop varietal diversity use to reduce pest and disease damage in agricultural ecosystems of Uganda. Agriculture, Ecosystems and Environment 157, 70– 86.

MusaNet, 2016. Summary of the global strategy for the conservation and use of genetic resources (R. Chase and B. Laliberté, compil.). Bioversity International, Montpellier, France, 27 pp.

Muyzer G, Brinkhoff T, Nübel U, Santegoeds C, Schafer H, Wawer C, 1997. Denaturing gradient gel electrophoresis (DGGE) in microbial ecology. In: Akkermans, A.D.L, van Elsas, J.D., de Bruijn, F.J., (Eds.) Molecular Microbial Ecology Manual. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp 1-27.

Nahalkova J, Fatehi J, Olivain C, Alabouvette C, 2008. Tomato root colonization by fluorescent-tagged pathogenic and protective strains of Fusarium oxysporum in

151

hydroponic culture differs from root colonization in soil. FEMS Microbiology Letters 286, 152-157.

Nel B, Steinberg C, Labuschagneand N, Viljoen A, 2006. Isolation and characterization of nonpathogenic Fusarium oxysporum isolates from the rhizosphere of healthy banana plants. Plant Pathology 55, 207–216.

Nelson PE, 1981. Fusarium: disease, biology, and taxonomy. University Park (Pa.). Pennsylvania State University Press. 457 pp.

Nielsen BO, Nielsen LB, 2004. Seasonal aspects of Sciarids emergence in arable land (Diptera: Sciaridae). Pedobiologia 48, 231-244.

Norgrove L, Hauser S, 2013. Black leaf streak disease and plantain fruit characteristics as affected by tree density and biomass management in a tropical agroforestry system. Agroforestry Systems 87, 349–354.

O’Donnell K, Kistler HC, Cigelnig E, Ploetz RC, 1998. Multiple evolutionary origins of the fungus causing Panama disease of banana: Concordant evidence from nuclear and mitochondrial gene genealogies. Proceedings of the National Academy of Sciences 95, 2044–2049.

O'Donnell K, Gueidan C, Sink S, Johnston PR, Crous PW, Glenn A, Riley R, Zitomer NC, Colyer P, Waalwijk C, et al., 2009. A two-locus DNA sequence database for typing plant and human pathogens within the Fusarium oxysporum species complex. Fungal Genetics and Biology 46, 936-948.

Okada T, Matsubara Y, 2012. Tolerance to Fusarium root rot and the changes in free amino acid contents in mycorrhizal asparagus plants. HortScience 47, 751 – 754.

Olivain C, Trouvelot S, Binet M-N, Cordier C, Pugin A, Alabouvette C, 2003. Colonization of flax roots and early physiological responses of flax cells inoculated with pathogenic and nonpathogenic strains of Fusarium oxysporum. Applied and Environmental Microbiology 69, 5453-5462.

Olivain C, Humbert C, Nahalkova J, Fatehi J, L'Haridon F, Alabouvette C, 2006. Colonization of tomato root by pathogenic and nonpathogenic Fusarium oxysporum strains inoculated together and separately into the soil. Applied and Environmental Microbiology 72, 1523-1531.

Omacini M, Eggers T, Bonkowski M, Gange AC, Jones TH, 2006. Leaf endophytes affect mycorrhizal status and growth of co-infected and neighbouring plants. Functional Ecology 20, 226-232.

Ordonez N, Seidl MF, Waalwijk C, Drenth A, Kilian A, Thomma BPHJ, Ploetz RC, Kema GHJ, 2015. Worse comes to worst: Bananas and Panama disease—when plant and pathogen clones meet. PLoS Pathogens 11 (11).

Ordonez N, García-Bastidas F, Laghari HB, Akkary MY, Harfouche EN, al Awar BN, Kema GHJ, 2016. First report of Fusarium oxysporum f. sp. cubense tropical race 4 causing Panama disease in Cavendish bananas in Pakistan and Lebanon. Plant Disease 100, 209.

152

Pan JJ, Baumgarten AM, May G, 2008. Effects of host plant environment and Ustilago maydis infection on the fungal endophyte community of maize (Zea mays). New Phytologist 178, 147–156.

Pantelides IS, Tjamos SE, Striglis IA, Chatzipavlidis I, Paplomatas EJ, 2009. Mode of action of a pathogenic Fusarium oxyspoprum strain against Verticillium dalhiae using Real Time QPCR analysis and biomarker transformation. Biological Control 50, 30–36.

Paparu P, Dubois T, Gold CS, Niere B, Adipala E, Coyne D, 2006. Colonisation pattern of nonpathogenic Fusarium oxysporum, a potential biological control agent, in roots and rhizomes of tissue cultured Musa plantlets. Annals of Applied Biology 149, 1–8.

Pattison AB, Wright CL, Kukulies TL, Molina AB, 2014. Ground cover management alters development of Fusarium wilt symptoms in Ducasse bananas. Australasian plant pathology 43, 465-476.

PBMH & PIF, 2006. Programa Brasileiro para a modernização da horicultura e produção integrada de frutas. Normas de Classificação de Banana. São Paulo: CEAGESP.

Pegg KG, Langdon PW, 1987. Fusarium wilt (Panama disease): a review. In: Persley, G. J., De Langhe, E. A. Banana and plantain breeding strategies. Proceedings of an International Workshop held at Cairns, Australia.

Peng HX, Sivasithamparama K, Turner DW, 1999. Chlamydospore germination and Fusarium wilt of banana plantlets in suppressive and conducive soils are affected by physical and chemical factors. Soil Biology and Biochemistry 31, 1363-1374.

Peng J, Zhang H, Chen F, Zhang X, Xie Y, Hou X, Li G, Pu J, 2014. Rapid and quantitative detection of Fusarium oxysporum f. sp. cubense race 4 in soil by real-time fluorescence loop-mediated isothermal amplification. Journal of Applied Microbiology 117, 1740-1749.

Pereira JCR, Pereira JR, Castro MEA, Gasparotto L, 2005. Ocorrência do Mal-do-Panamá em bananeira do subgrupo figo, em Piau, Minas Gerais. Fitopatologia brasileira 30, 554.

Perrier X, De Langhe E, Donohue M, Lentfer C, Vrydaghs L, Bakry F, Carreel F, Hippolyte I, Horry JP, Jenny C, Lebot V, Risterucci AM, Tomekpe K, Doutrelepont H, Ball T, Manwaring J, de Maret P, Denham T. (2011) Multidisciplinary perspectives on banana (Musa spp.) domestication. Proceedings of the National Academy of Sciences 108, 11311-11318.

Petrini O, 1991. Fungal endophytes of tree leaves. In: Andrews J.H., Hirano S.S. (eds) Microbial Ecology of Leaves. Brock/Springer Series in Contemporary Bioscience pp 179-197.

Ploetz RC, 1990. Fusarium wilt of Banana. APS Press, Minnesota, USA, 140 pp.

Ploetz RC, 2015. Management of Fusarium wilt of banana: A review with special reference to tropical race 4. Crop Protection 73, 7-15.

Ploetz RC, Pegg KG, 2000. Fusarium wilt. In: Diseases of Banana, Abaca and Enset.; Jones DR, editor. Wallingford, UK: CABI Publishing, pp 143-159.

153

Poland JA, Balint-Kurti PJ, Wisser RJ, Pratt RC, Nelson RJ, 2009. Shades of gray: the world of quantitative disease resistance. Trends in Plant Science 14, 21-29.

Postma J, Rattink H, 1991. Biological control of Fusarium wilt of carnation with a nonpathogenic isolate of Fusarium oxysporum. Canadian Journal of Botany 70, 1199- 1205.

Postma J, Schilder MT, Bloem J, van Leeuwen-Haagsma WK, 2008. Soil suppressiveness and functional diversity of the soil microflora in organic farming systems. Soil Biology and Biochemistry 40, 2394–2406.

Promusa, 2016. Morphology of the banana plant. Edited by: Vézina A, Van den Bergh I, Rouard M, http://www.promusa.org/tiki- index.php?page=Morphology+of+banana+plant, accessed on Dec 16th 2016.

Qamar A, Mysore KS, Senthil-Kumar M, 2015. Role of proline and pyrroline-5-carboxylate metabolism in plant defense against invading pathogens. Frontiers in Plant Science 6, 503.

Qi X, Guo L, Yang L, Huang J, 2013. Foatf1, a bZIP transcription factor of Fusarium oxysporum f. sp. cubense, is involved in pathogenesis by regulating the oxidative stress responses of Cavendish banana (Musa spp.). Physiological and Molecular Plant Pathology 84, 76-85.

Raaijmakers JM, Paulitz TC, Steinberg C, Alabouvette C, Moënne-Loccoz Y, 2009. The rhizosphere: a playground and battlefield for soilborne pathogens and beneficial microorganisms. Plant and Soil 321, 341–361.

Ramos B, Alves-Santos FM, Garcia-Sdnehez MA, Martin- Rodrigues N, Eslava AP, Diaz- Minguez JM, 2007. The gene coding for a new transcription factor (ftfl) of Fusarium oxysporum is only expressed during infection of common bean. Fungal Genetics and Biology 44, 864– 876.

Rep M, van der Does HC, Meijer M, van Wijk R, Houterman PM, Dekker HL, de Koster CG, Cornelissen BJC, 2004. A small, cysteine-rich protein secreted by Fusarium oxysporum during colonization of xylem vessels is required for I-3-mediated resistance in tomato. Molecular Microbiology 53, 1373–1383.

Roberts MR, 2007. Does GABA act as a signal in plants? Hints from molecular studies. Plant Signaling and Behavior 2, 408-409.

Robinson JC, Saúco VG, 2010. Bananas and Plantains, 2nd edition. Cab International, Oxfordshire, UK, 297 pp.

Rojas CM, Senthil-Kumar M,Tzin V, Mysore KS, 2014. Regulation of primary plant metabolism during plant-pathogen interactions and its contribution to plant defense. Frontiers in Plant Science 5, 17.

Sardos J, Perrier X, Dolezel J, Hribova E, Christelova P, Van den houwe I, Kilian A, Roux N, 2016. DArT whole genome profiling provides insights on the evolution and taxonomy of edible Banana (Musa spp.). Annals of Botany 118, 1269-1278.

Sasaki K, Nakahara K, Tanaka S, Shigyo M, Ito Si, 2015. Genetic and pathogenic variability of Fusarium oxysporum f. sp. cepae isolated from onion and welsh onion in Japan. Phytopathology 105, 525-532.

154

Saunders M, Kohn LM, 2009. Evidence for alteration of fungal endophyte community assembly by host defense compounds. New Phytologist 182, 229–238.

Saunders M, Glenn AE, Kohn LM, 2010. Exploring the evolutionary ecology of fungal endophytes in agricultural systems: using functional traits to reveal mechanisms in community processes. Evolutionary Applications 3, 525–537.

Scarlett K, Tesoriero L, Daniel R, Guest D, 2014. Sciarid and shore flies as aerial vectors of Fusarium oxysporum f. sp. cucumerinum in greenhouse cucumbers. Journal of Applied Entomology 138, 368–377.

Scher FM, Baker R, 1980. Mechanism of biological control in a Fusarium-suppressive soil. Phytopathology 70, 412-417.

Schippers B, Bakker AW, Bakker PAHM, 1987. Interactions of deleterious and beneficial rhizosphere microorganisms and the effect of cropping practices. Annual Review of Phytopathology 25, 339-358.

Schmidt SM, Houterman PM, Schreiver I, Ma L, Amyotte S, Chellappan B, Boeren S, Takken FLW, Rep M, 2013. MITEs in the promoters of effector genes allow prediction of novel virulence genes in Fusarium oxysporum. BMC Genomics 14, 119.

Schroth G, Krauss U, Gasparotto L, Duarte Aguilar JA, Vohland K, 2000. Pests and diseases in agroforestry systems of the humid tropics. Agroforestry Systems 50, 199–241.

Scott JC, McRoberts DN, Gordon TR, 2014. Colonization of lettuce cultivars and rotation crops by Fusarium oxysporum f. sp. lactucae, the cause of fusarium wilt of lettuce. Plant Pathology 63, 548-553.

Seifi HS, Van Bockhaven J, Angenon G, Höfte M, 2013. Glutamate metabolism in plant disease and defense: Friend or foe? Molecular Plant-Microbe Interactions 26, 475– 485.

Serra-Wittling C, Houot S, Alabouvette C, 1996. Increased soil suppressiveness to fusarium wilt of flax after addition of municipal solid waste compost. Soil Biology and Biochemistry 28, 1207–1214.

Shen Z, Ruan Y, Xue C, Zhong S, Li R, Shen Q, 2015. Soils naturally suppressive to banana Fusarium wilt disease harbor unique bacterial communities. Plant and Soil 393, 21– 33.

Silvar C, Merino F, Diaz J, 2009. Resistance in pepper plants induced by Fusarium oxysporum f. sp. lycopersici involve different defence-related genes. Plant biology 11, 68–74.

Simard SW, Durall DM, 2004. Mycorrhizal networks: a review of their extent, function, and importance. Canadian Journal of Botany 82, 1140–1165.

Simmonds NW, 1959. Bananas. London: Longmans, 466 pp.

Skovgaard K, BØdker L, Rosendahl S, 2002. Population structure and pathogenicity of members of the Fusarium oxysporum complex isolated from soil and root necrosis of pea (Pisum sativum L.). FEMS Microbiology Ecology 42, 367-374.

155

Smith SN, Snyder WC, 1971. Relationship of inoculum density and soil types to severity of Fusarium wilt of sweet potato. Phytopathology 61, 1049-1052.

Ssali RT, Kiggundu A, Lorenzen J, Karamura E, Tushemereirwe W, Viljoen A, 2013. Inheritance of resistance to Fusarium oxysporum f. sp. cubense race 1 in bananas. Euphytica 194, 425–430.

Staver C, Guharay F, Monterroso D, Muschler RG, 2001. Designing pest-suppressive multistrata perennial crop systems: shade-grown coffee in Central America. Agroforestry Systems 53, 151–170.

Steinberg C, Whipps JM, Wood DA, Fenlon J, Alabouvette C, 1999. Effects of nutritional sources on growth of one non-pathogenic strain and four strains of Fusarium oxysporum pathogenic on tomato. Mycological Research 103, 1210-1216.

Steinkellner S, Mammerler R, Vierheilig H, 2008. Germination of Fusarium oxysporum in root exudates from tomato plants challenged with different Fusarium oxysporum strains. European Journal of Plant Pathology 122, 395–401.

Stover RH, 1962. Fusarial Wilt (Panama Disease) of Bananas and Other Musa Species. CMI, Kew, Surrey, UK, 177 pp.

Stover RH, 1972. Banana, plantain and abaca diseaeses. Commonwealth mycological institute. Kew, UK, 316 pp.

Suarez-Capello C, Agama J, 2011. Relating crop damage levels on farm to crop varietal diversity measures by richness, evenness and divergence for banana in Ecuador. Damage, diversity and genetic vulnerability. The role of crop genetic diversity in the agricultural production system to reduce pest and disease damage. In: Jarvis, D.I., Fadda, C., De Santis, P., Thompson, J. (Eds.), Proceedings of an International Symposium. 15–17 February, 2011, Rabat, Morocco, Bioversity International, Rome.

Sun D, Lu X, Hu Y, Li W, Hong K, Mo Y, Cahill DM, Xie J, 2013. Methyl jasmonate induced defense responses increase resistance to Fusarium oxysporum f. sp. cubense race 4 in banana. Scientia Horticulturae 164, 484–491.

Sutherland R, Viljoen A, Myburg AA, van den Berg N, 2013. Pathogenicity associated genes in Fusarium oxysporum f. sp. cubense race 4. South African Journal of Science 109 (5/6).

Swarupa V, Ravishankar KV, Rekha A, 2014. Plant defense response against Fusarium oxysporum and strategies to develop tolerant genotypes in banana. Planta 239, 735– 751.

Swennen R, 2004. Banana, course notes plant production. Ghent Universty.51 pp.

Termorshuizen AJ, van Rijn E, van der Gaag DJ, Alabouvette C, Chen Y, Lagerlöfe J, Malandrakis AA, Paplomatas EJ, Rämerte B, Ryckeboer J, Steinberg C, Zmora- Nahum S, 2006. Suppressiveness of 18 composts against 7 pathosystems: Variability in pathogen response. Soil Biology and Biochemistry 38, 2461–2477.

Thaler JS, Owen B, Higgins VJ, 2004. The role of the jasmonate response in plant susceptibility to diverse pathogens with a range of lifestyles. Plant Physiology 135, 530-538.

156

Thatcher LF, Gardiner DM, Kazan K, Manners JM, 2012. A highly conserved effector in Fusarium oxysporum is required for full virulence on Arabidopsis. Molecular Plant– Microbe Interactions 25, 180–190.

Tixier P, Lavigne C, Alvarez S, Gauquier A, Blanchard M, Ripochea A, Acharda R, 2010. Model evaluation of cover crops, application to eleven species for banana cropping systems. European Journal of Agronomy 34, 53–61. van Dam P, Fokkens L, Schmidt SM, Linmans JHJ, Kistler HC, Ma LJ, Rep M, 2016. Effector profiles distinguish formae speciales of Fusarium oxysporum. Environmental Microbiology 18, 4087–4102.

Van den Berg N, Berger DK, Hein I, Birch PRJ, Wingfield MJ, Viljoen A, 2007. Tolerance in banana to Fusarium wilt is associated with early up-regulation of cell wall- strengthening genes in the roots. Molecular Plant Pathology 8, 333–341. van der Does HC, Lievens B, Claes L, Houterman PM, Cornelissen BJC, Rep M, 2008a. The presence of a virulence locus discriminates Fusarium oxysporum isolates causing tomato wilt from other isolates. Environmental Microbiology 10, 1475–1485. van der Does HC, Duyvesteijn RGE, Goltstein PM, van Schie CCN, Manders EMM, Cornelissen BJC, Rep M, 2008b. Expression of effector gene SIX1 of Fusarium oxysporum requires living plant cells. Fungal Genetics and Biology 45, 1257–1264.

Van Der Heijden MGA, Horton TR, 2009. Socialism in soil? The importance of mycorrhizal fungal networks for facilitation in natural ecosystems. Journal of Ecology 97, 1139– 1150.

Vander Molen GE, Beckman CH, Rodehorst E, 1987. The ultrastructure of tylose formation in resistant banana following inoculation with Fusarium oxysporum f.sp. cubense. Physiological and Molecular Plant Pathology 31, 185-200.

Van der Veken L, 2010. Potential of multipurpose intercrops for the management of pathogenic nematodes and beneficial arbuscular mycorrhizal fungi and root nodulating rhizobacteria in banana-based cropping systems. PhD thesis. KU Leuven, Belgium.

Vézina, A. (2015) The trap of extinction stories on bananas. http://www.promusa.org/blogpost424-The-trap-of-extinction-stories-on-bananas, accessed on May 15, 2017.

Vikili NG, 1965. Fusarium wilt resistance in seedlings and mature plants of Musa species. Phytopathology 55, 135– 140.

Vlaardingerbroek I, Beerens B, Rose L, Fokkens L, Cornelissen BJ, Rep M, 2016. Exchange of core chromosomes and horizontal transfer of lineage-specific chromosomes in Fusarium oxysporum. Environmental Microbiology 18, 3702-3713.

Waite BH, 1963. Wilt of Heliconia spp. caused by Fusarium oxysporum f. sp. cubense Race 3. Tropical Agriculture (Trinidad) 40, 299-305.

Wang Z, Zhang JB, Jia CH, Liu JH, Li YQ, Yin XM, Xu BY, Jin ZQ, 2012. De Novo characterization of the banana root transcriptome and analysis of gene expression under Fusarium oxysporum f. sp. cubense tropical race 4 infection. BMC Genomics 13, 650.

157

Waweru B, Turoop L, Kahangi E, Coyne D, Dubois T, 2014. Non-pathogenic Fusarium oxysporum endophytes provide field control of nematodes, improving yield of banana (Musa sp.). Biological control 74, 82-88.

Wehner J, Antunes PM, Powell JR, Caruso T, Rillig MC, 2011. Indigenous arbuscular mycorrhizal fungal assemblages protect grassland host plants from pathogens. PlosOne 6 (11).

Wieland G, Neumann R, Backhaus H, 2001. Variation of microbial communities in soil, rhizosphere and rhizoplane in response to crop species, soil type, and crop development. Applied Environmental Microbiology 67, 5849-5854.

Williams JS, Hall SA, Hawkesford MJ, Beale MH, Cooper RM, 2002. Elemental sulfur and thiol accumulation in tomato and defense against a fungal vascular pathogen. Plant Physiology 128, 150–159.

Wu HS, Liu DY, Ling N, Bao W, Ying RR, Ou YH, Huo ZH, Li YF, Shen QR, 2008a. Effects of vanillic acid on the growth and development of Fusarium oxysporum f. sp niveum. Allelopathy Journal 22, 111–121.

Wu HS, Raza W, Fan JQ, Sun YG, Bao W, Shen QR, 2008b. Cinnamic acid inhibits growth but stimulates production of pathogenesis factors by in vitro cultures of Fusarium oxysporum f. sp. niveum. Journal of Agricultural and Food Chemistry 56, 1316–1321.

Wuyts N, Lognay G, Verscheure M, Marlier M, De Waele D, Swennen R, 2007. Potential physical and chemical barriers to infection by the burrowing nematode Radopholus similis in roots of susceptible and resistant banana ( Musa spp.). Plant Pathology 56, 878–890.

Xu W, Wang Z, Wu F, 2015. Companion cropping with wheat increases resistance to Fusarium wilt in watermelon and the roles of root exudates in watermelon root growth. Physiological and Molecular Plant Pathology 90, 12-20.

Xue C, Penton CR, Shen Z, Zhang R, Huang Q, Li R, Ruan Y, Shen Q, 2015. Manipulating the banana rhizosphere microbiome for biological control of Panama disease. Scientific Reports 5, 11124.

Yadeta KA, Thomma BPHJ, 2013. The xylem as battleground for plant hosts and vascular wilt pathogens. Frontiers in Plant Science 4, 97.

Zhang H, Mallik A, Zeng RS, 2013a. Control of Panama disease of banana by rotating and intercropping with chinese chive (Allium Tuberosum Rottler): Role of plant volatiles. Journal of Chemical Ecology 39, 243–252.

Zhang Y, Lubberstedt T, Xu M, 2013b. The genetic and molecular basis of plant resistance to pathogens. Journal of Genetics and Genomics 40, 23-35.

158

Summary

Fusarium wilt of banana, better known as Panama disease, is one of the most challenging diseases of banana. The causal agent, the soil-borne fungus Fusarium oxysporum f. sp. cubense (Foc) induces vascular wilt and can survive for several decades in soil by the formation of chlamydospores. Foc is an assembly of genetically diverse F. oxysporum members pathogenic on banana. A subdivision in races is used based on the susceptible cultivar spectrum. Nowadays, Foc tropical race 4 (TR4) receives most attention, as it can attack Cavendish cultivars which are the most common cultivars used in export-oriented plantations. However, Foc race 1 is still a major pathogen for locally important cultivars, such as cultivar (cv.) Maçã (Silk, AAB) in Brazil.

In a farm in Brazil where banana was grown intercropped with trees and coffee, banana cv. Maçã was challenged by Foc race 1. A large variability in the disease was observed by the farmer. Via a greenhouse test, this variability could be attributed to differences in disease suppression in soil. Increasing the soil disease suppressive capacity would be a potential and desired option to manage Fusarium wilt in banana. Therefore, we investigated if disease suppressiveness could be correlated with particular soil abiotic properties, the plant community or the microbial soil community. Locations with higher disease suppression had a lower density of the susceptible cultivar, higher diversity in other banana cultivars, a lower ground cover with graminoids and a higher clay content and pH. The potential influence of banana cultivars and graminoids on the Foc and non-pathogenic F. oxysporum population was further investigated in this PhD thesis.

F. oxysporum isolates were obtained from soil, banana plants and graminoids of the farm. The collection comprised both Foc and non-pathogenic isolates. The isolates were molecularly characterized to study how the Foc population related to the non-pathogenic population to gain insight in the nature and origin of Foc. The main group of Foc had a different EF-1α and IGS marker gene sequence from the non-pathogens. Yet, three Foc isolates had an identical EF-1α/IGS sequence type as the largest group of non-pathogens. This resemblance suggests that these Foc isolates are of local origin. To study whether those isolates developed virulence on banana independently through coevolution with the host, or through chromosome transfer from other Foc isolates, we studied the presence and sequence of SIX genes, which are putative effector genes. An identical SIX1 sequence was encountered in all Foc isolates and was absent in the non-pathogenic isolates. The same

159

SIX1 sequence was found in Foc isolates from Australia. Based on SIX1, it can be hypothesized that the local Foc isolates acquired virulence through horizontal gene transfer. The possibility of transfer of virulence to non-pathogenic isolates has important consequences for the development of effective diagnostic tools and for plant breeders. Full genome sequencing is necessary to obtain a decisive explanation on the origin of the virulence in the Foc isolates that show resemblance to the local population.

Graminoids are known to be an excellent host for endophytic F. oxysporum. Therefore, they could contribute to the disease as symptomless carrier of Foc. Several graminoids were sampled on the field and F. oxysporum could be isolated from 50% of the plants. However, none of the isolates was pathogenic to banana. We investigated whether (1) Foc has a lower capacity to colonize symptomless hosts compared to non-pathogenic isolates, (2) Foc has a lower competitiveness in the presence of non-pathogenic F. oxysporum to colonize graminoids. Two graminoids were included in the test: Cyperus iria, a weed, and Brachiaria decumbens, a common cover crop. The capacity of Foc to colonize the roots was equal to that of non-pathogenic F. oxysporum and did not differ between the graminoids. However, when co-inoculated, non-pathogenic isolates colonized the roots better than Foc. The reduced competitiveness of Foc could not be explained by inferior growth of Foc compared to the non-pathogenic isolates. Our results show that it is important for the evaluation of plants as symptomless carrier of Foc to consider interactions between F. oxysporum isolates, and potentially also other micro-organisms.

The results of the field study suggested that banana cultivars with different level of resistance could have an impact on disease suppression in soil. Four cultivars with different level of resistance to Foc race 1 (Silk, very susceptible; Prata, moderately susceptible; Colatina Ouro and Dwarf Cavendish, both resistant) were studied on their interaction with Foc race 1 and non-pathogenic F. oxysporum. Both are important for the development of the disease: the first as pathogen and the latter as potential protective agent. Only the most susceptible cv. Silk stimulated Foc in soil. The level of root colonization by Foc corresponded to the level of susceptibility of the cultivars. Although moderately susceptible cv. Prata had the same internal symptoms as cv. Silk, its roots were less colonized and no stimulation in soil was observed. This result suggests an early root-located defense reaction of cv. Prata. In contrast to Foc, the root colonization of non-pathogenic F. oxysporum did not differ between the cultivars. After growth of cultivar Colatina Ouro, a larger population of non-pathogenic F. oxysporum in soil was observed. This cultivar could be discriminated from the others based on the amino acid content in the root exudates. Sugar levels in the root exudates did not differ between the cultivars. However, after inoculation with Foc, different shifts in sugar and

160

amino acid content of the root exudates of the cultivars were observed. These differential shifts are possibly a consequence of different defense reactions. Our results support the hypothesis that growing a mixture of bananas could moderate Fusarium wilt severity by restricting pathogen inoculum increases. Additional traits, such as the stimulation of non- pathogenic F. oxysporum in soil, could deliver an additional advantage. Field tests are necessary to evaluate the effect of cultivar mixtures.

The main objective of this thesis was to contribute in knowledge towards integrated management of Foc. We proposed how cultivar mixtures can have potential to moderate Fusarium wilt in banana. Although graminoids are possible symptomless carriers of Foc, it is unlikely that they have a negative impact on the disease. Finally, the proposed dynamics of the virulence of Foc has substantial importance for breeders and the development of diagnostic tools.

161

162

Samenvatting

Panamaziekte is een ernstige ziekte op banaan die veroorzaakt wordt door de bodemgebonden schimmel Fusarium oxysporum f. sp. cubense (Foc). Deze pathogeen kan gedurende decennia overleven in de bodem met behulp van chlamydosporen en veroorzaakt vasculaire verwelking bij banaan. Foc is de verzamelnaam voor een genetisch diverse groep F. oxysporum schimmels die pathogeen zijn op banaan. Foc wordt verder onderverdeeld in fysio’s gebaseerd op het spectrum vatbare variëteiten. Tegenwoordig gaat de meeste aandacht naar Foc fysio ‘tropical race 4’ (TR4). De meeste voorkomende cultivars in plantages, namelijk Cavendish cultivars, zijn vatbaar voor fysio TR4. Ondanks de grootste focus op TR 4, is Foc fysio 1 nog steeds een belangrijke pathogeen op banaan, maar dan op cultivars die belangrijk zijn voor de lokale markt, zoals cultivar (cv.) Maçã (Silk, AAB) in Brazilië.

Een Braziliaanse boerderij, waar banaan in mengteelt met bomen en koffie werd geteeld, had af te rekenen met Foc fysio 1 op banaan cv. Maçã. De boer had een grote variabiliteit in de ernst van de ziekte op de boerderij waargenomen. Die variabiliteit kon verklaard worden door verschillen in ziektebodemweerbaarheid met behulp van serre-proeven. Het bevorderen van de bodemweerbaarheid is een van de meest gegadigde mogelijkheden om panamaziekte onder controle te krijgen. Om inzicht te verwerven in factoren die kunnen bijdragen tot een verhoogde weerbaarheid, werd gezocht naar abiotische bodemfactoren, eigenschappen van de plantengemeenschap en microbieel bodemleven die eigen zijn aan de meest weerbare locaties op de boerderij. De meest weerbare locaties waren gekenmerkt door een hoger kleigehalte, hogere pH, geringere bodembedekking met grasachtigen, een lagere densiteit van de vatbare cultivar en hogere diversiteit van andere cultivars. De mogelijke invloed van grasachtigen en verschillende cultivars op Foc en niet-pathogene F. oxysporum populaties werd verder in dit doctoraat onder de loep genomen.

F. oxysporum werd geïsoleerd uit bodem, bananenplanten en grasachtigen van de boerderij. De collectie bestond uit zowel Foc als niet-pathogene F. oxysporum isolaten. Het vergelijken van de lokale niet-pathogene F. oxysporum populatie en Foc kan bijdragen tot inzicht in het ontstaan van pathogeniciteit van Foc. Met dit doel werd de moleculaire identiteit van de isolaten vergeleken op basis van twee kenmerkende genen: EF-1α en IGS. De grootste groep Foc isolaten had een EF-1α/IGS sequentie die afwijkend was van alle niet- pathogenen. Drie Foc isolaten hadden echter dezelfde EF-1α/IGS sequentie als de

163

voornaamste groep niet-pathogenen, wat een sterke aanwijzing vormt voor hun lokale afkomst. Het voorkomen en de sequentie van veronderstelde effector genen, namelijk SIX genen, werden vergeleken om inzicht te verwerven in het ontstaan van deze lokale pathogene isolaten. Pathogeniciteit kan verworven worden door een langdurige intieme relatie met de host (cfr co-evolutie) of door horizontale transfer van virulentiegenen tussen een Foc en niet-pathogeen isolaat. Het SIX1 gen werd enkel teruggevonden in Foc isolaten en had in alle Foc isolaten dezelfde sequentie. Dezelfde SIX1 sequentie was eerder al in Australische Foc isolaten beschreven. Het voorkomen van hetzelfde SIX1 gen in alle Foc isolaten doet vermoeden dat de lokale Foc isolaten virulentie genen hebben verkregen via uitwisseling met andere Foc isolaten. De mogelijke uitwisseling van de pathogeniciteitsgenen tussen F. oxysporum is van belang voor de ontwikkeling van efficiënte diagnostische methodes en voor de veredeling van resistentie in banaan. Om de transfer van pathogeniciteitsgenen te kunnen bevestigen, is sequenering van het volledige genoom van de betrokken isolaten onontbeerlijk.

Grasachtigen staan bekend als uitstekende gastplant voor endofytische F. oxysporum. Bijgevolg zouden zij kunnen bijdragen aan de ziekte als symptoomloze drager van Foc. Verscheidene grasachtigen werden verzameld op de boerderij. Uit de helft kon F. oxysporum geïsoleerd worden, maar geen enkel isolaat veroorzaakte symptomen in banaan. We hebben onderzocht of (1) Foc een beperktere capaciteit heeft om grasachtigen te koloniseren in vergelijking met niet-pathogene F. oxysporum, (2) Foc minder concurrentieel is in de aanwezigheid van niet-pathogene F. oxysporum om grasachtigen te koloniseren. Twee grasachtigen werden in de analyse ingesloten: Cyperus iria, een onkruid, en Brachiaria decumbens, een veel voorkomende bodembedekker. De capaciteit om de wortels te koloniseren was niet verschillend tussen Foc en de niet-pathogene isolaten. Wanneer Foc samen geïnoculeerd was met een niet-pathogeen isolaat, waren de wortels van de grasachtigen steevast minder door Foc gekoloniseerd dan door het niet-pathogene isolaat. Het lagere concurrentievermogen van Foc kon niet verklaard worden door inferieure groei. Onze resultaten tonen hoe onderlinge F. oxysporum interacties de kolonisatie van grasachtigen bepalen en hoe de bijdrage van grasachtigen aan de toename van Foc populaties vermoedelijk slechts beperkt is.

De resultaten van de veldstudie stellen een mogelijke rol voor verschillende bananencultivars in de opbouw van bodemweerbaarheid voorop. Vier cultivars met verschillende vatbaarheid voor Foc fysio 1 (Silk, zeer vatbaar; Prata, gemiddeld vatbaar; Colatina Ouro en Dwarf Cavendish, beide resistent) werden vergeleken voor hun invloed op Foc en niet-pathogene F. oxysporum. Beiden kunnen de ontwikkeling van de ziekte beïnvloeden: Foc als pathogeen

164

en niet-pathogene F. oxysporum als potentiële biocontrole organismen. Enkel de zeer vatbare cv. Silk stimuleerde Foc in de bodem. De wortelkolonisatie door Foc was overeenkomstig met de vatbaarheid van de cultivar. Ondanks gelijkaardige interne symptomen bij cv. Silk en cv. Prata, stimuleerde cv. Prata Foc niet in de bodem en was de wortelkolonisatie beperkt. Dit zou kunnen wijzen op een vroege en in de wortels gelokaliseerde afweerreactie bij cv. Prata. In tegenstelling tot Foc, was de wortelkolonisatie door de niet-pathogene isolaten niet verschillend tussen de cultivars. In de bodem werd een grotere niet-pathogene populatie gevonden na groei van cv. Colatina Ouro dan bij de andere cultivars. Deze cultivar onderscheidde zich ook het duidelijkst van de andere gebaseerd op het gehalte aan aminozuren in de wortel-exudaten. Het suikergehalte in de exudaten verschilde niet tussen de cultivars. Na inoculatie echter, waren de veranderingen in suikergehaltes verschillend. De veranderingen in aminozuur en suikerconcentratie in de exudaten na inoculatie zouden het gevolg van diverse afweerreacties kunnen zijn. Onze resultaten onderbouwen de hypothese dat het groeien van cultivar mengsels de ziekte zou kunnen beperken door het inperken van de Foc toename. Andere eigenschappen, zoals het stimuleren van niet-pathogene F. oxyporum, kunnen een bijkomende bijdrage leveren aan het bevorderden van de bodemweerbaarheid.

Dit doctoraatsproject beoogde bij te dragen aan de geïntegreerde bestrijding van Fusarium verwelking in banaan door extra kennis aan te brengen over ecologische interacties van Foc met verschillende elementen van de bananenteelt. We stellen voor dat het mengen van bananenvariëteiten kan bijdragen in de beheersing van de ziekte. Daarnaast toonden we aan hoe grasachtigen waarschijnlijk maar beperkt bijdragen als symptoomloze dragers van Foc door onderlinge competitie binnen F. oxysporum. Tot slot, de voorgestelde uitwisseling van pathogeniciteit tussen F. oxysporum isolaten is belangrijk voor de ontwikkeling van diagnostische methoden en voor de veredeling van banaan.

165

166

Curriculum Vitae

PERSONAL DETAILS

Surname: Deltour First name: Pauline Address: Maurice Verdoncklaan 91, 9050 Gentbrugge Cell phone: 0032489888606 Email: [email protected] Date of birth: 07/08/1989 Place of birth: Lokeren Nationality: Belgian

RESEARCH EXPERIENCE

PhD, Ghent University (2012 - present) Subject: “Penetrating the network of Fusarium oxysporum populations, banana cultivars and graminoids:Towards holistic Fusarium wilt management in banana ” Field: Phytopathology Promotor: Prof. dr. ir. Monica Höfte Co-promotor: dr. ir. Soraya de Carvalho França

Inagro vzw: applied agricultural research (September 2017 – present) Research coordinator of the cluster organic agriculture

EDUCATION

Master in Bioscience Engineering: Agricultural Sciences, Ghent University (2010-2012) Thesis: “Beheersing van Rhizoctonia solani op sla door het bevorderen van de bodemweerbaarheid”. Promotor: Prof. dr. ir. Monica Höfte Co-promotor: dr. ir. Soraya de Carvalho França

Bachelor in Bioscience Engineering, Ghent University (2007-2010)

High school: Greek-Mathemathics, Sint-Lodewijkscollege Lokeren (2001-2007)

167

ADDITIONAL TRAININGS

Project Management, Ghent University (2016) Communication skills, Ghent University (2016) Authentic Networking, Ghent University (2016) Academic Posters, Ghent University (2016) Arctic Microbiology, University of Akureyri, Iceland (2014) Practical Portuguese 1, Ghent University (2012) Information Cycle BTC, Belgian Technical Cooperation (2011)

EXPERIENCE ABROAD

Research stay at the Federal University of Viçosa, Brazil (May 2013 - December 2013) Supervision: Prof. dr. ir. Olinto Liparini Pereira and Prof. dr. ir. Irene Cardoso

TEACHING EXPERIENCE

Practical courses on Tropical Crop protection (2013-2016)

SUPERVISION OF UNDERGRADUATE STUDENTS

Ellen Velkeneers (2012 - 2013) Suppression of Fusarium oxysporum f. sp. cubense on banana in an agroforestry system in Brazil. Thesis to obtain the degree of Master of Science of Nutrition and Rural Development: Tropical Agriculture.

Filip Snauwaert (2013 – 2014) Karakteristatie van Fusarium oxysporum geassocieerd met banaan en grasachtigen in een agroforestry systeem in Brazilië. Thesis to obtain the Master degree in Bioscience Engineering: Agricultural Sciences.

Lisa Heyman (2014 – 2015) Non-pathogenic Fusarium oxysporum populations as drivers of soil suppressiveness to Fusarium wilt in banana. Thesis to obtain the Master degree in Bioscience Engineering: Forest and Nature Management.

168

PUBLICATION LIST

Peer reviewed

Deltour P, França SC, Pereira OL, Cardoso I, De Neve S, Debode J, Höfte M (2017) Disease suppressiveness to Fusarium wilt of banana in an agroforestry system: Influence of soil characteristics and plant community. Agriculture, Ecosystems and Environment 239, 173– 181.

Deltour P, França SC, Heyman L, Pereira OL, Höfte M (2017) Comparative analysis of pathogenic and non-pathogenic Fusarium oxysporum populations associated with banana on a farm in Minas Gerais, Brazil. Plant Pathology, in press.

CONFERENCE CONTRIBUTIONS

66th International Symposium on Crop Protection, Ghent, Belgium May 2014 (Oral presentation) Deltour P, Velkeneers E, França SC, Pereira OL, Cardoso I, Höfte M. Soil spots suppressive to banana Fusarium wilt in an agroforestry system in Brazil.

IXth International Symposium on Banana / ISHS-ProMusa symposium, Brisbane, Australia, August 2014 (Oral presentation) Deltour P, Velkeneers E, França SC, Pereira OL, Cardoso I, Höfte M. Induction of suppressive soil spots to Fusarium wilt in banana in an agroforestry system in Brazil.

68th International Symposium on Crop Protection, Ghent, Belgium, May 2016 (Poster presentation) Deltour P, França SC, Höfte M. Harmful or beneficial? Graminoids as reservoir of pathogenic and non-pathogenic Fusarium oxysporum.

Tropentag 2016, Vienna, Austria, September 2016 (Oral presentation) Deltour P, França SC, Höfte M. The influence of banana cultivars on pathogenic and non- pathogenic Fusarium oxysporum.

Xth International Symposium on Banana / ISHS-ProMusa symposium, Montpellier, France, October 2016 (Poster presentation) Deltour P, França SC, Höfte M. Harmful or beneficial? Graminoids as reservoir of pathogenic and non-pathogenic Fusarium oxysporum.

Xth International Symposium on Banana / ISHS-ProMusa symposium, Montpellier, France, October 2016 (Poster and short oral presentation) Deltour P, França SC, Höfte M. The influence of different banana cultivars on pathogenic and non-pathogenic Fusarium oxysporum.

169

OTHER SCIENTIFIC COMMUNICATION

Participation in the Troca de Sabers 2013, an extension event on the Federal University of Viçosa between farmer, academics and students. Organization of a game in which participants could get into the world of microorganisms and Fusarium wilt.

Participation in the ‘Wetenschapsbattle’, 10 May 2016: Scientists present their research on an elementary school and the children vote for the best presentation.

AWARDS

Bayer award for best master thesis about plant sciences (2012)

170